EPA-650/4-74-051
FEASIBILITY OF METRAC SYSTEM
FOR REGIONAL AIR POLLUTION STUDY
by
Robert W. Johnson, Kenneth S. Gage,
William H. Jasperson, Robert C. Rust, and Richard K. Kirchner
Control Data Corporation
Research Division
8100 South 34th Avenue
Minneapolis, Minnesota 55440
Contract No. 68-02-0760
ROAPNo. 26AAI
Program Element No . 1AA003
EPA Project Officer: Thomas J. Lemmons
Meteorology Laboratory
National Environmental Research Center
Research Triangle Park, North Carolina 27711
Prepared for
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON, D.C. 20460
August 1974
-------
EPA REVIEW NOTICE
This report has been reviewed by the National Environmental Research
Center - Research Triangle Park, Office of Research and Development,
EPA, and approved for publication. Approval does not signify that the
contents necessarily reflect the views and policies of the Environmental
Protection Agency, nor does mention of trade names or commercial
products constitute endorsement or recommendation for use.
This document is available to the public for sale through the National
Technical Information Service, Springfield, Virginia 22161.
11
-------
TABLE OF CONTENTS
I. CONCLUSIONS 1
II. RECOMMENDATIONS 1
III. INTRODUCTION 2
IV. SYSTEM DESCRIPTION 3
A. METRAC System Description 3
1. Balloon Transmitter Function 5
2. Reference Transmitter Site Function... 7
3. METRAC Receiver Site Function 8
4. Command Site Function 9
5. Minneapolis Site Function 10
6. Computation Function 10
B. Data Format 10
1. Receiver Sample Format 10
2. Cassette Tape Format 13
3. 6600 Tape Format 13
V. HARDWARE 15
A. Balloon Transmitter 15
B. Reference Transmitter 18
C. Receiver 26
1. Radio Frequency Section 26
2. Analog Processing Section 37
3. Data Acquisition Section 46
D. Receiver Checkout Unit 51
-------
E. Command Site Console 51
1. Data Acquisition 59
2. Data Monitor 65
3. Reference Transmitter Control 68
F. Minneapolis Site 70
VI. METRAC POSITION DETERMINATION 73
A. Discussion of the Problem 73
B. Effect of Geometry 77
C. Sources of Error 78
D. METRAC Solution 79
E. Software 82
1. Software Input 82
2. Program METRAC 83
VII. MINNEAPOLIS FIELD TEST 86
A. METRAC Deployment for the Minneapolis Test 87
B. Preliminary System Tests 90
C. Wind Profile Comparison Tests 92
D. Special Capabilities of the METRAC System 106
VIII. BIBLIOGRAPHY 114
-------
iii
LIST OF FIGURES
FIGURE PAGE
1. Metrac System 6
2. Receiver Sample Format 12
3. Cassette Tape Format 14
4. Photograph of Packaged Balloon Transmitter 16
5. Block Diagram of Transmitter RF Section 17
6. Balloon Transmitter Performance (prototype) 19
7. Performance (temperature compensated) 20
8. Block Diagram Reference Transmitter Site Equipment 22
9. Photograph of Tripler and Power Amplifier Modules 25
10. Photograph of Transmitter Equipment Enclosure 27
11. METRAC Receiver Site Equipment 28
12. RF Section - METRAC Receiver System 30
13. RF Circuit Specification 33
14. Analog Processing Circuits 40
15. Photograph of Receiving Antenna and Cavity 42
16. Photograph of Receiver Front Panel 43
17. Photograph of Receiver Interior 44
18. Photograph of RF Module 45
19. Data Acquisition Logic Block Diagram 49
20. Photograph of Receiver Checkout Unit 52
21. Block Diagram METRAC Receiver Checkout Unit 53
22. Block Diagram of Command Site Equipment 55
23. Photograph of Command Site Front Panel 58
24. Photograph of Front Panel Construction 60
25. Photograph of Command Site Equipment Rear View 61
26. Block Diagram CS Data Acquisition logic 63
27. Command Site Data Monitor Block Diagram 67
28. Block Diagram CMD Site Reference Transmitter Control 69
29. Block Diagram of Data Link 71
30. 6600 Tape Generation and Processing 72
31. Hyperbolas Formed by Two Receivers 74
-------
LV
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
Intersecting Hyperbolas Formed by Two Receiver Pairs
Sample Balloon Trajectory for Triangular Receiver Array
Three Examples of Height Error Versus Height for Receiver
Position Inaccuracy
Flow Diagram of METRAC Software Package
Map of Installations for Twin Cities Test
X Versus Y Plot of Walk on Penthouse Roof of Radisson Hotel
with Data Plotted at 1 Second Intervals
XY Trajectories for Selected Test Flights
Comparison of 60 Second Wind Profiles for Flight MF3
Comparison of 60 Second Wind Profiles for Flight MF4
Comparison of 60 Second Wind Profiles for Flight MF7
Comparison of 20 Second Wind Profiles for Flight MF3
Comparison of 20 Second Wind Profiles for Flight MF4
Comparison of 20 second Wind Profiles for Flight MF7
Wind Profiles for Flight MF2
Wind Profiles for Flight MF5
Wind Profiles (20 sec) for Flight MF5
Wind Profile for Flight MF2
Comparison of 60 Second and 30 Second Profiles
Comparison of 15 Second and 60 Second Profiles (MF2)
Comparison of 15 Second and 60 Second Profiles (MF7)
One Second Samples for Flight MF4 Showing Oscillation of
Suspended Transmitter
Vertical Velocity of Balloon Versus Time and Height
76
80
81
84
89
91
93
94
95
96
97
98
99
103
104
105
107
108
109
110
111
113
-------
-1-
I. CONCLUSIONS
TM
METRAC is an especially attractive technique to collect wind soundings of great
accuracy and high-resolution in support of large field programs. Its feasibility
has been adequately demonstrated in an urban environment during the Minneapolis
field test. The accuracy and resolution apparent from the results of the field
test suggest that the prototype system has a performance capability comparable to a
radar tracking system. With minor modifications to the hardware and some
development of software the present system can be employed to obtain wind sound-
ings in support of the Regional Air Pollution Study Program. A further develop-
ment effort will be required to incorporate temperature and humidity sensing
capability into the system. Multiple tracking can be provided by duplicating
parts of the system.
II. RECOMMENDATIONS
Because of a limitation of funding it was not possible to carry out the radar
comparison test originally proposed for this feasibility study. A careful radar
comparison test should still be made in order to further evaluate the accuracy
and resolution capabilities of METRAC. Such a test should provide the information
required to devise an optimal method to calibrate the system.
In order to realize the full potential of the METRAC system as an atmospheric
probe, a further development effort will be required. Temperature and humidity
sensors and modulation circuits should be added to the transmitter. Detector
circuits for demodulation should be built into the receiver. Multiple tracking
should be the goal of another development effort. This can be achieved by
duplicating parts of the system.
-------
-2-
III. INTRODUCTION
TM
METRAC is a ground-based radio location system which employs the Doppler principle
to track an inexpensive, lightweight, expendable transmitter. The transmitter can
be attached to a vertically rising balloon in order to obtain an accurate, high-
resolution, sounding of the atmospheric wind field. Alternatively, the transmitter
can be attached to a horizontally free-floating balloon in order to determine
atmospheric trajectories.
The METRAC system was originally conceived during the summer of 1965 as an
economical means to obtain highly accurate wind data for air pollution studies.
Since then, it has been under development by the Research Division of Control
Data Corporation. Limited scale tests of the tracking system were performed in
1966 and again in 1969 verifying the principles of operation.
In 1969 the METRAC system was evaluated in competition with several other track-
ing systems by the MITRE Corporation (1969) under contract to ESSA. MITRE's
effort under Project SESAME (System Engineering Study for Atmospheric Measurements
and Equipments) provided information to assist Weather Bureau decision makers in
selecting the basic vertical atmospheric sounding system for operation in the
decade beginning in the early 1970's. For operational use MITRE recommended the
NAVAID technique to the Weather Bureau since it had already undergone considerable
development. The METRAC system was considered potentially superior in performance
to the NAVAID systems but was not recommended because of the necessity of a costly
research and development effort to implement the system.
Partly as a result of MITRE's report, Stanford Research Institute (1972) recommend-
ed the use of the METRAC system to the Environmental Protection Agency for its
Regional Air Pollution Study (RAPS) Program. The present paper is the result of
an evaluation of the feasibility of using the METRAC system for the RAPS program.
For this reason results presented here are concerned primarily with low-level winds.
-------
-3-
This report documents the results of a study conducted by Control Data Corporation
to test the feasibility of employing the METRAC approach for collecting upper air
data in support of the Regional Air Pollution Study (RAPS) Program being conducted
by the Environmental Protection Agency in St. Louis.
The major portion of this report contains an engineering description of the
prototype METRAC system that was fabricated for this feasibility study. The
report concludes with a presentation of results from a field test conducted in
Minneapolis by Control Data Corporation. This test consisted of a comparison of
wind profiles obtained from METRAC with profiles obtained by simultaneously
tracking the same balloon with a theodolite and a rawinsonde system.
IV. SYSTEM DESCRIPTION
A. Metrac System Description
The theoretical basis of METRAC and the electronics required to implement this
system are not highly complicated. However, this system is a departure from
commonly used tracking systems and its principles are sometimes confused. There-
fore, a descriptive explanation of overall system operation is presented here.
The METRAC system is based on the Doppler shift of a moving transmitter. Although
the physical principles are well known, only the recent availability of low-cost
digital components and UHF-VHF transistors have permitted an economically feasible
electronic design. METRAC uses omnidirectional antennas for both transmitting
and receiving and does not require mechanical or electronic scanning. This
eliminates the elaborate pedestal and drive assemblies and the limited capability
under extreme environmental conditions associated with dish antenna tracking systems.
-------
-4-
The basic elements of METRAC are an airborne transmitter and several receiving
stations having known positions. As the transmitter moves in space, the frequency
at each receiver equals the transmitted frequency plus a Doppler frequency shift,
which is a linear function of the velocity of the transmitter. Because the true
transmitted frequency may not be known, this Doppler shift cannot be determined
from only the data at one receiver. However, the data from any pair of receivers
permits a determination of the difference of received frequencies. Since these
receivers are at rest with respect to each other, the frequency difference equals
the difference of the Doppler shift associated with the receiver pair. This
Doppler difference is the only data required to determine the transmitter
position relative to the receivers.
The Doppler shift is a fundamental element of the METRAC system. The Doppler
principle asserts that the received frequency of a signal is higher than the
transmitted frequency as the transmitter moves toward the receiver. The received
frequency is lower than the transmitted frequency as the transmitter moves away
from the receiver. The received signals consist of the transmitted frequency,
f , and a Doppler shift, d., where the subscript i refers to the receiver. The
received frequency is then f + d.. The transmitted frequency is not exactly
known because the oscillator drift may be on the order of 10 KHz. This drift
is significantly larger than the Doppler shift, which is less than 100 Hz for
balloon tracking. The frequency difference for any receiver pair (i and j) is
then
tf = f\ - fj = (fT + d.) - (fT + d..) = dt - d.. (1)
Thus, the frequency difference is equal to the Doppler difference.
Since it is anticipated that the frequency difference for any receiver pair will
be found using counters at each receiver site, the reception process can be
simplified considerably by using a reference transmitter whose signal reaches
-------
-5-
all receivers in the system. The reference frequency f is then effectively
R
subtracted from the received frequency indicated above, such that f. becomes
simply f + d. - f . The frequency difference Af between each receiver pair
T i R
is unchanged, and remains equal to the Doppler difference.
The integrated Doppler difference associated with each pair of receivers is
directly proportional to the slant range difference from the transmitter to each
receiver. A known slant range difference determines a hyperbolic line of position
on which the transmitter is located. The receivers are the foci of this hyper-
boloid. The data from three independent receiver pairs (four receivers) deter-
mines the transmitter position in space.
The electronics required to implement the system consists of an inexpensive
balloon-borne transmitter, four or more receivers and a control station. These
superheterodyne receivers are less sophisticated than good commercial FM receivers.
The control station is used to record the Doppler data to determine the transmitter
position.
A METRAC position determining system is illustrated in Figure 1. Variations in
the position of a mobile transmitter are translated by the system into a computer
generated output which gives detailed trajectory information. The functions of
the system components are outlined in the following text.
1. Balloon Transmitter Function
The mobile transmitter, hereafter referred to as a balloon transmitter, transmits
an unmodulated radio frequency carrier. Motion of this transmitter relative to
a network of receivers will produce a shift in the received signal frequencies.
The resulting Doppler shift information is recovered, combined with position data
describing the location of receiver sites and the balloon transmitter at a known
time, and then processed to generate the position information.
The transmitter operates at a frequency of 403 MHz. The operating frequency is
crystal controlled to minimize the required receiving bandwidth and to permit
the simultaneous tracking of multiple balloons. The effective radiated power
-------
-6-
£
M
OT
O
8
S
£
u
Cri
Ctf
RECEIVING SITE 2
13
O
B
n
w
u
§
H
^3
CO
o
M
En
-------
-7-
output is 400 mW which provides an operating range in excess of 25 km. Vertical
polarization is used to enhance multipath performance. The transmitter was de-
signed as a reasonable cost, lightweight, expendable item. Total weight is less
than 300 gms and the unpackaged volume is less than 24 cubic inches.
The transmitter package is subjected to a wide range of temperature in flights
extending from the surface to the tropopause; the electronic circuits are designed
to provide essentially constant characteristics over a -30 C to +60 C range.
Packaging techniques extend this region to -55 C for flights of duration less
than one-half hour.
2. Reference Transmitter Site Function
The reference transmitter site supplies two system functions. First, the reference
transmitter eliminates the requirement that receiver sites measure the absolute
frequency of the balloon signal. Instead, the receiver sites measure the
difference in frequency between the reference and balloon signals. Second, the
reference transmitter is phase modulated with a sine wave. All receiver sites
recover this modulation to derive a constant frequency which permits improved
sampling performance. This improvement results when the reference modulation
signal is subtracted from the multiplied receiver signal. The use of multipli-
cation increases the basic resolution of the Doppler measurement, and the offset
by the reference modulation reduces the nominal receiver count rate F , thus
DO
reducing sampling error.
The reference transmitter site provides a power output of 5 watts effective
radiated power. This level insures an adequate reference transmitter signal at
receiver sites such that difference frequency information will not be degraded
due to reference channel noise and such that the reference transmitter modulation
will have an adequate signal to noise ratio. To enhance the received signal, the
antenna should have a line of sight path to all receivers. This requirement was
satisfied in the Minneapolis test by mounting the reference antenna at the top
of a 600 foot building (IDS Center).
-------
-8-
System operation requires that the frequency of the reference transmitter must
be adjusted to compensate for variations in the operating frequency of the
balloon transmitter caused by changes in temperature and battery voltage. Such
frequency variations are usually slow, long period, monotonic changes. This
adjustment may be accomplished by manual control at the reference transmitter
site or by remote control from the command site console. The use of remote
control requires a voice bandwidth link from the command site. The reference
transmitter frequency may be adjusted + .005%; this is sufficient to accommodate
the frequency changes caused by nominal and environmental balloon transmitter
variations.
3. METRAC Receiver Site Function
A network of METRAC receiver sites is required to recover Doppler information.
The METRAC receiver sites receive transmissions from the balloon and reference
transmitters. The difference in frequency is extracted, then multiplied by
eight to increase system position resolution. The multiplied frequency difference
is offset by subtracting a constant frequency derived from reference transmitter
modulation. This signal (FD ) is accumulated in a digital counter. Status
information describing the condition of receiver circuits and receiver signals
is continually produced. Receipt of a sample command from the command site
results in the transmission of a sample of the contents of the FD accumulator,
the status information, and receiver ID.
The receiver is tuned by sweeping to acquire the reference signal. The command
to acquire the reference is transmitted by the command site, and a single sweep
of the receiver results, with a sweep duration of about 30 seconds. Once the
receiver has acquired the reference, it operates with automatic frequency control
to keep the reference acquired.
Another command permits clearing the FD accumulator to a count of zero.
Commands from the command site and data to the command site are exchanged over
leased telephone lines. The data are exchanged as eight bit characters.
Asynchronous transmission of characters using 1200 baud modems is employed.
The present METRAC receiver is constructed to operate over a 0 to 70 C temperature
range.
-------
-9-
4. Command Site Function
The command site is the common data collection point in the METRAC system. The
command site acquires data from all receiver sites. Data is recorded as
sequential records on a cassette tape unit. After termination of a mission the
cassette produced is played back to generate a CDC 6600 compatible tape.
The command site also provides control of the receiver site and reference trans-
mitter site equipment. Commands transmitted to receiver sites permit acquiring
the reference signal and zeroing of accumulators. Reference transmitter carrier
may be turned on or off; frequency may be controlled manually or automatically.
System data output is monitored with a command site data monitor. The data
displayed provides reference transmitter frequency control information and
permits an evaluation of system performance. All system data is communicated
over leased voice bandwidth telephone lines.
The data acquisition function requires the sampling of data at all receiver sites,
the transmission of this data to the command site, and the formation of a sample
record containing data from all sites. Ideally the FD accumulators in all
METRAC receivers would be sampled simultaneously. This condition is approximated
by simultaneously transmitting sample commands to all receiver sites. The time
difference in receiver sampling results in a short term error of not more than
one count per sample. No accumulative count errors from sampling uncertainty
can occur.
The receiver sites transmit their data upon receipt of the sample command.
Eight characters are used to represent station ID, the FD accumulator, status
information, and space for sensor information. The command site stores this
data in buffer registers. When data from all receivers has been stored, the
registers are read out sequentially to form a single sample record of data from
all receivers. This record is stored on a digital cassette recorder.
-------
-10-
5. Minneapolis Site Function
The Minneapolis site generates a CDC 6600 compatible tape from the data recorded
on the command site cassette unit. The command site tape can be transmitted
over telephone lines to the Minneapolis site to produce an identical cassette
tape. This cassette is played back, reformatted, and recorded on a 6600
compatible tape transport.
6. Computation Function
The tape containing receiver site frequency difference information and additional
data describing receiver site locations and the balloon transmitter location
at launch provide the input for trajectory computation. The computed output
contains three dimensional position coordinates and velocities as a function of
time. Additional output indicates system errors and the path taken to bypass
these problem areas.
B. Data Format
The data sampled at each receiver is made up of eight 8-bit characters, and is
appropriately termed a Receiver Sample. By simultaneous sampling of all receivers
a complete record can be compiled containing data from each receiver for a given
sample. This record is termed a Sample Record.
1. Receiver Sample Format
The receiver sample includes the following data:
Number of Bits
1. Receiver Identification Number 4
2. FD Accumulator Counter 28
o
3. Status 8
4. Sensor 24
-------
-11-
The data is then divided into eight bit characters as shown in Figure 2. All
data (except the ID) in the receiver sample is encoded in a BCD format. Thus,
one decimal digit requires four bits of data such that two digits are packed
into one eight bit character for transmission. The receiver ID is programmed
for each receiver unit using hexidecimal representation. Seven decimal digits
are used to represent the FD accumulator count. This requires twenty-eight
bits of data. Status information is allocated eight bit locations, although
only six are presently used. Twenty-four bits have been reserved for sensor
data; this permits the transmission of two three-digit BCD numbers. These
numbers might typically represent the temperature and pressure measurements
from balloon-borne sensors.
The status information transmitted is defined in the following manner:
Bit Number
1
2
3
4
5
6
7,8
Name
Reference Acquired
AFC Mode
Tracking Filter
Mode
Power Status
FD > FMAX
FD < FMIN
SPARES
Definition
Logic "1" indicates receiver is
detecting the reference signal.
Logic "1" indicates AFC circuit
is in a track mode.
Logic "1" indicates 2.8 KHz
tracking filter is locked to an
acceptable signal.
Logic "1" indicates power is on.
Logic "1" indicates instantaneous
frequency is greater than limit.
Logic "1" indicates instantaneous
frequency is less than limit.
Bit positions 1-4 are normally at a logic "1". Bit positions 5-8 are normally
"0". Deviation from this condition indicates data may be in error.
-------
-12-
C8
C7
C6
C5
C4
C3
C2
Cl
First Character on Tape
Cl
First Four Bits
of Accumulator
Counter. (FD BCD)
o
Rcvr Identification Number
C2
Third FD BCD
o
Second FD BCD
o
C3
C4
C5
27
2°
Fifth FD BCD Fourth FD BCD
o o
27
2°
Seventh FD BCD Sixth FD BCD
o o
27
26
25
*- %''
Spare
24
23
22
21
2°
1 1
L Reference Acquired
LAFC Mode
^-Tracking Filter
"- Power
LFD > FD MAX
L- FD < FD MIN
Status
C6, C7, C8 all zeros
FIGURE 2: RECEIVER SAMPLE FORMAT
-------
-13-
2. Cassette Tape Format
The cassette tape contains a series of Sample Records; one record is written for
each sample period and each Sample Record contains a Receiver Sample for each
network receiver site. The cassette is written incrementally with eight-bit
characters as shown in Figure 3. The first character on tape is Cl (character 1)
for the Receiver Sample of receiver #1. C2 through C8 are written, followed by
Cl of receiver #2's data. After a complete Sample Record of 56 characters is
written, a sync word containing 8 ones is written. A total of 57 characters
(8 char/receiver x 7 receivers + one sync) is thus written for each Sample Record.
The receiver record is derived as follows:
Bits = 448 Bits (Data) + 8 Bits (Sync) = 456 Bits/Sample Record
Sample Record Sample Record Sample Record
Character = 8 Bits
8 Characters = Receiver Sample
7 Receiver Samples = Sample Record
Therefore:
Sample Record = 8 Char x 7 Receivers + 1 Char Sync = 57 Characters
Receiver
3. 6600 Tape Format
The 6600 compatible tapes are made by grouping characters in groups of threes or
24 bits. Reformatting the data from three eight bit characters to four 6-bit
frames gives the necessary conversion for 6600 processing. The format logic in
the tape interface further divides the data to appear as follows:
3 Characters = 24 Bits
Frame = 6 Bits
Sub-Character = 4 Frames
19 Sub-Characters = Record
-------
-14-
1. CHARACTER
1
2
3
4
5
6
7
8
«KL-Bit
(8-Bits)
RECEIVER SAMPLE
1
^. T- -1
2 3
<-l-Chara
(8 Ch
4
ctei
San
arac
5 6
ters)
7 8
_^
SAMPLE RECORD
1
2
3
4
5
6
7
*Rcvr #1 H
SYNC
t 7 Rcvr Samoles _ *
tAll Ones
(8 Bits)
and Sync
FIGURE 3: CASSETTE TAPE FORMAT
-------
-15-
V. HARDWARE
A. Balloon Transmitter
The balloon transmitter design goal was to produce a compact lightweight, and
reasonable cost expendable package. This package provides a crystal controlled
radio frequency signal at 403.0 MHz with an effective radiated power level of
400 mW. The frequency is held to a tolerance of less than + .005% and power is
maintained essentially constant over a temperature range exceeding -30 to +60 C.
A picture of the packaged balloon transmitter appears in Figure 4. The package
contains a battery pack, a R.F. section, and an antenna. The battery pack contains
nine 1.5V cells wired in series with a battery connector. Transmitter power is
enabled by attaching a mating connector from the R.F. section. The battery pack
and R.F. section are assembled, tested, then encapsulated in a polyurethane foam.
Finally, a water repellant cardboard covering is attached to the units.
A block diagram of the radio frequency section appears in Figure 5. Three function-
al blocks are present. The crystal oscillator used operates at 134.33 MHz; a
seventh overtone crystal determines operating frequency. This oscillator is a
commercially available module. The tripler stage contains a bipolar transistor
operated as a grounded emitter Class C frequency multiplier. The base circuit is
tuned to 134.33MHz and the collector circuit is tuned to 403.00MHz. The collector
circuit provides impedance matching to the power amplifier stage containing a
single bipolar transistor Class C amplifier. A double tuned circuit couples the
amplifier to the load. More than 200 mW of load power is obtained at 50 ohms using
a 13.5 VDC supply.
Bias to the tripler and amplifier stages is temperature compensated. The R.F.
section is constructed on a teflon impregnated circuit board. Transmission lines
etched on this material form the inductors for the resonant circuits. Air variable
capacitors permit tuning these circuits.
The radiating structure for the transmetter is formed by a flexible quarter-wave
element and the ground plane of the R.F. section and battery pack. The transmitter
is normally operated in a vertical orientation, giving the pattern of a vertical
dipole. Circuit pads are provided such that horizontal quarter-wave elements can
be mounted. With capacitive connection to the transmitter output terminal, an
elliptical polarization can be obtained, permitting operation with a transmitter
directly above a receiving antenna.
-------
-------
-17-
CXO
X3
P.A.
FREQUENCY
403.000 MHz +.0057o
POWER
ERP w 400
SIZE
2x2x8"
(Excluding Antenna)
WEIGHT
< 300 grams
(Including Batteries)
TEMPERATURE
-30°C to +60°C
ANTENNA
Dipole
SPURIOUS OUTPUT
< -36 dB relative to
fundamental
B.C. POWER
13V (§80 ma («1W)
FIGURE 5: BLOCK DIAGRAM OF TRANSMITTER RF SECTION
-------
-18-
Figure 6 indicates the performance of a prototype METRAC balloon transmitter.
The power output is strongly dependent on temperature below 0 C. The design was
modified by incorporating temperature compensation of bias voltage. The resulting
design was again tested versus temperature; results are shown in Figure 7. Power
output is essentially constant to -30 C.
Tests were performed in the METRAC laboratory to verify the performance of the
balloon transmitter package with temperature and voltage variations. The package
would operate with less than 1 dB of power degradation and acceptable frequency
change for one-half hour after immersion in a -50 C cold box.
During several of the,Minneapolis tests, an additional foam package was con-
structed of two inch polystyrene. This added insulation gave extended battery
operation during the high level radiosonde comparison flights.
A shortcoming of the transmitter design is that only marginal drive is available
at the power amplifier stage. This causes tuning to be somewhat critical. A
further potential problem with the transmitter is the crystal itself. During
tethered balloon tests it was found that sudden accelerations could cause rela-
tively large changes in transmitter frequency. This effect was minimized by
attaching the transmitter to the balloon using a cord attached to the antenna.
A period of faulty operation was noted for many flights. Each transmitter would
experience a rapid frequency shift of several hundred cycles. This problem lasted
perhaps one minute and was associated with balloon positions near the tropopause.
B. Reference Transmitter
The reference transmitter site performs two system functions. A) It provides a
radio frequency signal located 2.8 KHz below the balloon transmitter frequency
which is available at all receiver sites. By forming FD, the difference between
-------
-19-
En
g
-------
PS
H 6
co
(2 o
H PH
5 |~~ vO
1 csl CM
I
I
-20-
in -d-
CM CM
+ +
1 1
1 1
CO
+
1 ,
1
H U
^
I
M
U
e
PS
w
u
H
S
w
o
p^
o*
O
-------
-21-
balloon and reference frequencies, the Doppler difference formulation is made
independent of receiver local oscillator frequencies. B) Transmitted reference
modulation is available at all receiver sites. In the present system, this
modulation permits a reduction in FD accumulator rate to reduce sampling error.
As the balloon transmitter frequency varies due to temperature and battery
voltage changes, the reference transmitter must be adjusted to nominal 2.8 KHz
offset to maintain normal receiver operation. Transmitted power level must
provide sufficient receiver site signal so that: 1) the FD signal/noise ratio
is not significantly degraded by mixing in the reference channel noise, and 2)
the detected reference modulation has an adequate signal/noise ratio.
The present reference transmitter site generates a 5 W (ERP) signal. The
frequency may be varied + .005% from 403 MHz. The signal is phase modulated
with a crystal oscillator derived 267 Hz sine wave. Figure 8 is a block diagram
of the equipment.
The transmitter is provided with local and remote control of transmitter frequency
and carrier (on/off). Locally, frequency is controlled with a ten-turn
potentiometer. Carrier is turned on or off with a switch. The remote control
mode of operation requires a telephone line link to the command site. Two narrow-
band FSK tone channels decode command site control transmissions, while carrier
on/off is controlled by a relay closure. A frequency to voltage converter trans-
lates tone modulation to develop the transmitter frequency control voltage.
The reference transmitter radio frequency chain contains a VCXO module, a phase
modulator module, a tripler module, a power amplifier module, a tuned cavity, a
transmission line, and an antenna. Transmitter frequency is determined by the
-------
-22-
PHONE
LINE
FROM
CMD
SITE
403 MHz
5W ERP
TRIPLER
MODULE
P.A.
MODULE
CAVTTY CA.RTNET
VCXO
TONE
RCVR
2
PHASE
MODULATOR
FREQ.
MULT.
SINE
WAVE
SOURCE
LOCAL/
REMOTE
SWITCH
TONE
RCVR
1
F/V
CONV.
0
134.333
MHz
13 VDC
13V
xA/V^ - *
P.S
LOCAL FREQ. CONT.
117 VAC
FIGURE 8: BLOCK DIAGRAM REFERENCE TRANSMITTER SITE EQUIPMENT
-------
-23-
control voltage input to the VCXO module. The output of the voltage controlled
crystal oscillator is phase modulated and then frequency multiplied to the module
output frequency of 134.333 MHz. The frequency tripler module provides 403. MHz
drive to the power amplifier module. The signal is filtered by the cavity,
carried by the transmission line to the antenna, and radiated.
The VCXO module contains a fundamental mode crystal oscillator. The frequency
is varied + .005% for a 0-10V input voltage change which varies the bias of a
voltage variable capacitance diode. The phase modulator is connected between
this oscillator and two stages of frequency multiplication. These tripler stages
provide a 134.333 MHz RF signal. A diplexing circuit combines DC power with
this signal for transmission to the tripler module.
The phase modulator is designed to produce essentially ideal modulation char-
acteristics. This is necessary to guarantee that modulation products will not
cause interference at the balloon signal frequencies. The modulator peak phase
deviation can be varied. The present METRAC system operates with a peak deviation
of .3 radian.
The tripler module receives RF at 134.333 MHz and DC power through a 50 ohm
transmission line. The two inputs are separated and the RF is tripled to 403 MHz.
The DC powers both tripler and power amplifier modules. The frequency tripler
function is realized using a modified METRAC balloon transmitter.
The power amplifier module provides power gain at 403 MHz. Greater than four
watts of module output may be obtained. The power gain is varied with a
potentiometer. A hybrid amplifier circuit, low pass filter, and power control
circuit are contained within the module.
-------
-24-
The tuned cavity is inserted in the system to reduce transmitter spurious output
and to minimize interaction with neighboring R.F. sources. No intermodulation
problem was detected during Minneapolis testing.
The transmission line used is a 50 ohm foam dielectric cable. Due to the power
level, a braided shield cable was acceptable.
The antenna used is an omnidirectional gain type antenna, transmitting a
vertically polarized signal. The antenna provides 8.9 dB gain referred to an
isotropic source. The antenna gain permits obtaining the 5 watt ERP with
moderate transmitter power and transmission line loss.
The modulating signal at the reference transmitter site is generated by the sine
wave source card (SWS). The demodulation circuits at the receiving site have a
narrow bandwidth (< 10 Hz) for noise performance; the modulating frequency must
therefore be quite stable. The moderate separation of balloon and reference
frequencies requires low harmonic content modulating signals. These two
conditions are met by controlling the signal frequency with a crystal oscillator.
Two cascaded active filters provide an essentially sinusoidal voltage output.
The frequency produced may be changed by altering the division factor in a
multistage frequency divider.
The reference transmitter electronic equipment is housed in two cabinets. A
cavity cabinet contains the tripler module, power amplifier module and cavity.
The VCXO module, phase modulator module, and power supply are housed in an
instrument enclosure. A locked outer housing contains this enclosure and the
remote control equipment. 117 VAC power is required to operate the system. A
photograph of the tripler and power amplifier modules (Figure 9) shows the con-
struction used. The cases used were designed to house CATV equipment in an outdoor
-------
-------
-26-
environment. This fact, coupled with the remote DC power capability mentioned,
would permit mounting this equipment on a tower near the antenna. An additional
photograph (Figure 10) shows the equipment enclosure used to mount the VCXO,
phase modulator, and power supply. The modules are the same as those used for
METRAC receiver construction.
C. Receiver
The METRAC receiver site monitors balloon reference transmitter signals and
processes these signals to provide a digital message (receiver sample) which
contains the Doppler difference information. The receiver sample further contains
status information describing the condition of the receiver site equipment. It
could also include information from balloon borne sensors. The receiver sample
is encoded for transmission via voice grade telephone circuits. Commands
initiated by the command site and communicated via these circuits are decoded to
control the time of sampling of data, acquisition of the reference signal, and
zeroing of receiver accumulators.
Figure 11 shows the three functional elements present at each receiver site;
the RF, analog processing, and data acquisition sections. All components except
the antenna, tower, transmission line, and cavity are housed within the receiver
cabinet. The receiver sample can be monitored at the receiver site by connecting
a receiver checkout unit to a test connector mounted on the front panel of the
receiver cabinet.
1. Radio Frequency Section
The Radio Frequency Section contains: 1) antenna, tower, and transmission line;
2) the tuned cavity; 3) the RF circuits of the METRAC receiver.
-------
-------
-28-
g,
pSl
CO
---. co
O O
EH 55
< O O
Q CJ O
S5
O
M
< EH
EH M
-< co
Q M
O1
O
CO
co
O £H
O M
hJ t3
-------
-29-
Antenna, Tower and Transmission Line
The METRAC receiving antenna is a Right Hand Circularly Polarized (R.H.C.P.)
antenna developed by the Andrew Corporation for reception of satellite
transmissions. The antenna possesses a hemispherically shaped response pattern.
Pattern gain is 0 dBiC at the horizon, + .SdBiC directly overhead, and -7 dBiC at
45 below the horizon. The antennas measure 11 ini
A 12 inch stem below the antenna permits mounting.
45 below the horizon. The antennas measure 11 inches square by 15 inches high.
If balloon transmitters are provided with circularly polarized antennas, tracking
can be accomplished directly above a receiving site. Towers used in the METRAC
system support the antenna and provide elevation above rooftops, etc., to improve
multipath rejection for high elevation angles.
The transmission line used in METRAC is Andrew Corporation FHJ2-50A 50 ohm foam
Heliax ; this is a solid conductor sheathed, foam dielectric cable. The
attenuation at 400 MHz is 2.5 dB/100 feet.
The tuned cavity is a high-Q resonant circuit which provides additional selectivity
preceding the RF circuits. This was required in our urban test to minimize effects
due to local T.V. transmissions, UHF mobile radio transmissions, etc. The units
were purchased from a commercial supplier and modified in the Research Division
mechanical lab.
RF Circuits
Figure 12 illustrates the RF circuits in the METRAC receiver. These circuits form
a triple conversion superheterodyne receiver with automatic frequency control of
tuning. Balloon and reference signals at 403 MHz are filtered to remove inter-
fering sources, translated to 455 KHz, separated by highly selective filters,
and amplified in limiting amplifiers. The balloon out and reference out signals
are the inputs to the Analog Processing Section which contains the Doppler
difference information; their difference in frequency is FD.
-------
-30-
I
00
Pi
I
H
U
W
CA
O
H
fa
-------
-31-
The receiver is tuned by varying the frequency of the local oscillator injection
at the converter. The VCXO, which determines local oscillator frequency, is
controlled by VAFC from the AFC card. Two modes of AFC operation are provided.
Receipt of an "Acquire Reference" command from the command site or a receiver panel
switch initiates the "Sweep" mode. Upon command the receiver initializes a
frequency search from 402.980 to 403.020 MHz. Presence of an acceptable signal
at reference out changes the "Reference Acquired" status bit to a logic 1 and
sets the AFC mode to "AFC Track". The reference out signal now serves as the
controlled variable in the AFC loop as the VCXO is tuned to maintain a constant
reference out frequency.
AFC card status is transmitted by the Data Acquisition Section to the command
site. "AFC Track" = logic 1 indicates the track mode. "Reference Acquired" =
logic 1 indicates an acceptable reference signal is present. The AFC card also
recovers reference transmitter modulation for use by the Analog Processing
Section.
Several factors affected the design of the RF circuits:
1) METRAC is a differential system which requires the simultaneous reception
of two R.F. signals.
2) METRAC utilizes non-directional receiving antennas.
3) The METRAC equipment is intended to operate unattended for relatively long
periods of time.
4) In an urban environment the equipment would likely share antenna sites
with other VHF/UHF communications equipment.
The differential nature of the system imposes the most stringent requirements.
The balloon transmitter signal power detected at a receiver site located within
one hundred feet of a balloon launch point could vary by > 80 dB. The reference
-------
-32-
signal will be fixed in amplitude but at a much lower power level, perhaps at
the minimum useable level. To maintain linearity, minimal power gain £» 30 dB)
precedes the filters which separate the two signals.
Commercially available mechanical filters designed for single sideband radio
applications were selected and used in METRAC for signal separation. These
filters provide sufficient bandwidth (1.7 KHz) for the passage of balloon and
reference modulation. Center frequency separation is small, 2.8 KHz, providing
a moderate FD and permitting sampling with a voice bandwidth channel. A penalty
associated with the frequency separation is that the balloon and reference modula-
tion must be controlled to prevent splatter into the adjacent frequency range.
Low index (|3 ~«3 to -5) phase modulation is compatible with this requirement.
The dynamic range requirement associated with the balloon signal, coupled with
the expectation that angle modulation would be used in the balloon package to
maximize carrier power and minimize modulator power requirements, led to the
use of limiting amplifiers following the mechanical filters.
Omnidirectional receiving antennas operating in an urban environment (with
attendant interference sources) motivated the decision to build a highly
selective receiver. The requirement for long-term stability suggested that a
triple conversion approach would be desirable. This approach also permits
utilization of nearly optimum conversions in the mixing process. Image frequency
rejection exceeds 70 dB in each conversion, while spurious product rejection
approaches 70 dB. Figure 13 shows RF Circuit Specification.
Converter
The converter accepts balloon and reference signals at 403 MHz, filters and
amplifies these signals, and translates them to 35.6 MHz. A conversion power
-------
-33-
FIGURE 13. RF CIRCUIT SPECIFICATION
SIGNAL INPUT:
Impedance = 50 ohms
Noise Figure = 3-5 dB
Frequency Tuning Range =
402.980 to 403.020 MHz
(403 MHz +.005%)
SIGNAL OUTPUT:
IF Bandwidth = 1.7 KHz Nominal
Output Level = 1 V p.p Nominal
Output Frequency = 453.6 KHz (Reference)
456.4 KHz (Balloon)
DYNAMIC RANGE:
15 dB S/N at -120 dBm
1 dB Compression -30 dBm
Balloon/Reference Differential < 50 dB
TUNING:
Track Mode - Reference out maintained at
453.6 KHz +40 Hz
Sweep Mode - Searches 402.980 - 403.020 MHz
SPURIOUS RESPONSE:
Image Rejection « 70 dB
-------
-34-
gain of 25 dB is obtained with an input noise figure of 3 to 5 dB. Two stages of
R.F. gain precede the mixer. The mixer Local oscillator injection at 367.4 MHz is
developed by a transistor tripler driven by the 122.4666 MHz output of the VCXO
module.
Image frequency rejection is in excess of 70 dB. Spurious responses caused by the
non-fundamental local oscillator injection are down ~70 dB with a tuned cavity in
the input signal line. One dB of signal compression occurs at an input signal
level -35 dBm.
The converter used is a modified commercial unit developed by Vanguard Laboratories.
Modifications included the addition of shielding to improve tuning stability, the
change to accept drive from the VCXO, and minor mechanical changes to provide
compatibility with the receiver packaging system.
35.6 MHz IF
This IF is driven by the converter; signals are filtered and translated to 4.5 MHz.
A power gain of about 5 dB is obtained, along with a noise figure of 18 dB and an
image frequency rejection of 70 dB. The module contains a 1.5 MHz bandwidth three
pole filter, a MOSFET mixer, and a 31.1 MHz crystal oscillator. The units were
designed, assembled, and tested by Control Data.
4.5 MHz IF
The signal input at 4.5 MHz is filtered and translated to 455 KHz. Low impedance
outputs are provided to drive the balloon and reference IF strips. A power
gain of 5 dB is obtained, along with a noise figure of 12 dB and an image frequency
rejection of 70 dB. The filter function provides an RF bandwidth of 200 KHz. An
MOSFET mixer driven by a crystal oscillator performs the frequency translation.
-------
-35-
Balloon I.F.
Signals at 456.4 KHz + 850 pass through a highly selective mechanical filter.
Cascaded amplifier and limiter stages provide a nominal output signal of IV
p-p for receiver inputs in excess of -120 dBm. Module power gain is ~ 107 dB.
The mechanical filter used provides the rejection of the reference signal required
for system operation. The units provide > 65 dB rejection. The amplifier is
constructed using 3 voltage gain stages with broadband crystal filter interstage
coupling. The first stage contains a low noise discrete transistor. The second
stage uses an integrated circuit limiting amplifier. The third stage is also an
I.C. limiter; an emitter follower drives the output connector to provide isolation
and a low impedance output.
These units were designed, constructed, and tested at Control Data. Satisfactory
operation was obtained but with some difficulty due to the fact that about 110 dB
of power gain is contained within the package.
Reference I.F.
These circuits are similar to the balloon I.F. except that the signal frequency
is 453.6 KHz + 850 Hz.
AFC Card
The AFC card controls receiver tuning by varying the VAFC tuning voltage to the
VCXO module. The card operates in either a track or sweep mode; the sweep mode
being associated with the acquisition of the reference signal. An "Acquire
Reference" command decoded by the data acquisition section (or initiated by a
receiver panel switch) sets the AFC card in the "Sweep" mode. VAFC then increases
from 0 to 10V, tuning the receiver from 402.980 to 403.020 MHz. Detection of an
-------
-36-
acceptable reference signal causes a "Reference Acquired" indication. This
transition initiates the "Track" mode of operation. The reference out signal
frequency becomes the controlled variable in an AFC loop formed by the AFC card,
the VCXO module, the converter, 35.6 MHz and 4.5 MHz I.F.'s, and the Reference
I.F. VAFC is varied to maintain the reference out signal at 453.6 KHz. The
"Reference Acquired" and "AFC Track" indications are two of the status bits
transmitted by the data acquisition section to the command site.
The AFC circuitry is designed to be stable and accurate, maintaining the reference
out frequency within + 50 Hz of nominal with transmitter frequency variations of
+ 20 KHz. The reference output signal at 453.6 KHz is the signal input to a
mixer located on the AFC card. A 450 KHz crystal oscillator provides the local
oscillator signal. The mixer output at 3.6 KHz drives two circuits: a narrow-
band tone detector and a frequency discriminator. The narrowband tone detector
initiates a "Reference Acquired" indication when a signal is present at 3.6 KHz
+ 180 Hz. The frequency discriminator produces a voltage proportional to signal
frequency. This voltage is filtered to drive an error amplifier. When operating
in the "Track" mode this error voltage is amplified to produce VAFC. VAFC in
the Sweep mode is produced by a ramp voltage generator. An FET Analog multiplexer
selects the source which develops VAFC.
The voltage output of the frequency discriminator also recovers frequency modula-
tion for use by the analog processing section.
VCXO
The VCXO module is the voltage variable frequency source in the METRAC receiver.
VAFC variation from 0-10V tunes the receiver from 402.980 MHz to 403.020 MHz.
The module output is a 122.466 MHz sine wave at a power level of 10 mW.
-------
-37-
To directly obtain frequency variations of + .005% from a crystal oscillator it
is necessary to use a fundamental crystal. This fundamental signal is multiplied
by cascaded frequency multipliers to obtain the desired output frequency. The
fundamental crystal oscillator is tuned with a voltage variable capacitance
diode. Oscillator output is doubled, then tripled to the output frequency. A
double tuned output filter circuit is used to minimize spurious receiver responses
due to local oscillator products.
2. Analog Processing Section
The analog processing section accepts the constant level balloon and reference
output signals. The difference frequency (FD ~ 2.8 KHz) is formed, then filtered
to reduce the noise bandwidth to ~ 400 Hz. The FD signal is then multiplied by
a factor of eight resulting in MXFD at ~ 22.4 KHz. Finally, the MXFD signal is
offset by a constant factor to produce FD at ~ 5 KHz. The FD signal which
contains multiplied FD information then increments the data acquisition section
FD accumulator. An FD monitor circuit generates status error information when
receiver S/N is not adequate for reliable operation.
The bandwidth obtained with the mechanical filters is excessive in terms of the
signal bandwidth required by the received R.F. carriers. Further, a METRAC
system simulation indicated a multiplication of Doppler difference information
would be required to obtain resolutions comparable with initial estimates. This
requisite multiplication also indicates that a reduction in bandwidth is in order.
The 2.8 KHz tracking filter was inserted to provide the desired reduction in
effective bandwidth. The filter's 3 dB bandwidth was set at 200 Hz. The un-
certainty in the frequency difference suggested using a phase locked loop to
provide the filter function. The loop would track the difference frequency as
the quantity varied. The tracking filter also contains a sweep circuit and a
lock detector. These components were added to permit acquisition of the signal
as the tracking bandwidth was insufficient to permit loop capture for difference
-------
-38-
frequency variations of + 800 Hz. The lock detector determines if the filter is
locked to the input signal.
A monitor circuit was also designed into the analog processing circuits to
monitor received signal quality by examining the signal present at
the output of the FD difference circuit. Because of the limiting I.F. amplifiers,
amplitude variations have been removed from the signals; therefore, the frequency
characteristics only are monitored. Frequency deviations exceeding
+1.4 KHz are detected to set error indicating latches. The phase deviation
o
associated with this indication is about + 90 . An error indication does not
guarantee an error in accumulation has occurred, but indicates the received S/N
ratio is in question.
The FD multiplier circuit was designed to provide a variable multiplication
factor. The factor of eight presently implemented is a compromise between in-
creasing the quantization resolution and decreasing the signal to noise ratio at
the accumulator. The circuit was designed to have a signal bandwidth sufficient
to pass FD signals at 2 through 3.6 KHz. The output frequency for a nominal FD
is ~ 22.4 KHz.
The 22.4 KHz rate resulting from the addition of the FD multiplier circuit required
the addition of an FD offsetting circuit. This circuit subtracts a constant
frequency from the multiplied FD. The Doppler information is retained, but the
rate of accumulation is reduced which results in reduced sampling errors for a
fixed sampling time uncertainty. The offsetting frequency used must be identical
at all receiver sites to prevent the accumulation of a count difference between
receivers. Accordingly the offsetting frequency is derived from reference trans-
mitter modulation. Low frequency (~ 268 Hz) sinusoidal modulation is compatible
with system requirements to minimize interaction between the reference and balloon
signals. Detected modulation is filtered to reduce frequency uncertainty and
-------
-39-
then multiplied to ~ 17 KHz. Subtracting this constant from 22.4 KHz resultings
in a offset frequency difference (FD
reduction in sampling error results.
in a offset frequency difference (FD ) signal at 5 KHz. A factor of five
Figure 14 shows the block diagram of the Analog Processing Section; constant
level balloon out (456.4 KHz) and reference out (453.6 KHz) signals are
differenced by the FD Difference Circuit to produce FD (^2.18 KHz). The FD
signal is filtered by the 2.8 KHz phase lock loop tracking filter, then multi-
plied by eight in the FD multiplier circuit, to produce MXFD (22.4 KHz). The
FD difference circuit differences MXFD and a 17 KHz signal derived from
reference modulation. The output FD containing the multiplied Doppler difference
information increments the Data Acquisition Section FD accumulator.
Additional Analog Processing Section outputs are status bits. The FD monitor
continuously examines the output of the FD difference circuit. An instantaneous
deviation in FD beyond preset limits sets error latches. Status bit "FD > FMAX"
indicates that FD exceeded the upper frequency limit during a sample interval
whereas the "FD < FMIN" status bit indicates that FD was less than the lower
frequency limit.
The P.L.L. tracking filter status is indicated by the "Filter Track" status bit;
logic 1 indicates that the FD signal is acceptable and is being tracked by the
loop. Logic 0 indicates the tracking filter center frequency is being swept to
locate an acceptable signal.
The FD difference circuit contains a linear multiplier to mix the balloon and
reference signals and a bandpass filter to recover the difference frequency
component. The bandpass output drives a limiter and the 2.8 KHz P.L.L. tracking
filter. The limiter output drives the FD monitor which generates the "FD > FMAX"
and "FD < FMIN" status bits.
-------
'40-
^
>o
o
k
CM M
tt
*
CJ
cv, N
S^
O
I-l w
U hJ
t
V.
Pd
EH
H
a
d
<
is
&.
,1
j
4
1
i
00 t
CN'E
i >
O
t,
4
f
1
f
t
4
<
^
[ STATUS
4
J t
:\
j
3 J
k 4
1
< 2
n
*\ N
3 6
t ^
g
Q g
K M
t
;
H
-1
/
3
-1
\
o-t
N t
S3
0
O
co
c.
1-
nJ fc<
hJ t-
P-i <
r
1
3
ES
§
I
»
P
t
< ^d"
J 10
3^
t
s
^ P Q C
3 hj ca c
CM E ;
S Pn * H
k A
N
S3 C
§ 1
s
?) *"
4
<
§
M
S 1
1 !
a
O CO
t-i pa
co
.A.
3
» c.
i t
^ i
2 w g
< t t
a
q
^
-(
Ej
J
»
C
k
y
<
n
t
i
M
M
M
co
m
W
§
H ^
Q 8
h-J
k
-sj"
Pi
8
H
NOH03S
-------
-41-
The 2.8KHz tracking filter was required to permit FD multiplication without a
drastic increase in the system signal/noise requirement. The filter contains
a phase locked loop providing a bandwidth of 200Hz, a lock monitor, and a sweep
circuit. The lock monitor controls the mode of loop operation; if frequency lock
is maintained between the loop and the FD signal, a "Filter Track" indication
is produced. If frequency lock is not maintained, the mode changes to sweep.
The sweep circuit varies the center frequency of the P.L.L.; when frequency lock
is established, the mode reverts to track.
The Analog Processing Section functions are contained on three circuit cards;
the FD difference circuit, limiter, and 2.8KHz tracking filter are on one card.
The FD monitor and FD multiplier are on a second card, while the P.L.L. tracking
filter, multiplier, and FD difference circuit form a third card.
The various components required at a METRAC receiver site are illustrated in a
series of photographs. Figure 15 is a picture of the METRAC receiving antenna
and tuned cavity. Figure 16 shows the METRAC receiver front panel. Front
panel connectors are used for telephone line connections, the RF input cable,
the receiver checkout unit, and for AC power.
The receiver was mechanically designed to permit rapid access to all components;
Figure 17 illustrates the interior of the receiver. The RF circuits are housed
in modules developed by CDC. Signal connections are made using co-axial cable
and mating R.F. connectors while power is supplied to the modules through pins
mating with chassis connectors. The analog processing circuits (not shown)
and the data acquisition logic are mounted on printed circuit cards and plugged
into a card rack; spare slots are provided for the addition of circuitry to
recover sensor information. All major components can be removed in a few
moments.
The modules used for the RF circuits are shown in Figure 18, in which the outer
shield has been removed. Rapid access is provided.
-------
FIGURE 15. PHOTOGRAPH OF
RECEIVING ANTENNA
AND CAVITY
-------
-------
-------
-
FIGURE 18. PHOTOGRAPH OF
RF MODULE
-------
-46-
3. Data Acquisition Section
The design of the Data Acquisition Section was affected by requirements to
minimize the possibility of introducing errors due to telephone line noise
or heterodyne error, to accommodate receiver status information, to permit
the addition of sensor information, and to allow the remote control of receiver
functions.
To reduce the probability of introducing errors in the Doppler difference form-
ulation, the accumulator is located at the receiver site. The accumulator runs
continuously and is sampled to provide data. Instantaneous noise bursts on the
telephone line can only introduce errors during a particular sample interval;
the accumulator is unaffected. Second, the data are transmitted digitally;
this provides some protection as certain transmission errors may be detected by
monitoring character parity, etc. Third, command transmissions to the receiver
sites are encoded with redundancy to minimize the probability of the receiver
responding to a false transmission.
The selection of digital data transmission improves flexibility since the data
message was readily expanded to include status bits, and space for sensor informa-
tion could be included. The digital system also provides the medium for command
transmissions. Such transmissions are encoded for reliability as discussed in
the following text.
The basic data rate chosen was 1200 baud to provide reasonable resolution in the
sampling process. Sampling time uncertainty of about + .25ms is achieved.
Ideally the FD accumulators in all METRA.C receivers would be simultaneously
sampled. This condition is approximated by simultaneously transmitting sample
commands to all receiver sites. No accumulative differential timing errors can
occur. The sampling process is nonideal because of differential delays in the
pulse transmission paths and uncertainties in pulse detection time due to infinite
bandwidth and signal/noise ratio. Differential delay was measured to be less than
10ms while the uncertainty in detection time was minimized by using 1200 baud
modems to transmit data.
-------
-47-
The METRAC receiver data acquisition section performs two system functions:
1) Commands initiated by the command site are received and decoded
to control receiver site circuitry, and
2) Receiver site data are accumulated, sampled, then transmitted
to the command site.
The data acquisition section is capable of decoding four commands. Presently,
three commands are used. Received commands zero the FD accumulator which
o
generates the multiplied Doppler signal data, place the receiver AFC circuits
in a sweep mode to acquire the reference, or sample the receiver data. A
fourth command could control receiver power status, etc. The commands are de-
modulated by a modem to a serial data format, converted from serial to parallel
using a UART module, and decoded in a command decoder. Redundancy is used to
improve command transmission reliability.
Acquisition of data at each receiver site is initiated by loading a buffer
register upon receipt of a sample command; the resultant receiver sample presently
contains accumulated Doppler frequency information and digital status bits
describing the condition of receiver circuitry and signals. Provision has been
made to incorporate data from up to two sensors in the data message.
The buffer register contains four bits of receiver identification, seven digits
of FD accumulator data, and eight bits of status information. The data are
o
converted from parallel to serial using a UART, and transmitted by a modem
connected to leased telephone lines. The serial data is received by the command
site and processed to form a sequential record on a cassette data recorder.
-------
-48-
The METRAC Data Acquisition Section logic at each receiver site consists of three
logic cards and a modem; these appear in Figure 19. These logic cards are
labeled with the following mnemonics!
RBD-1 --- ACCUM./REG.
RBD-2 --- UART/CMD-DEC.
RBD-3 --- RCVR Timing
RBD-1
This card accumulates the multiplied, offset, frequency difference information
and provides a buffer register for sampled receiver data. The FD signal pulses
from the Analog Processing Section are counted in a seven decade BCD counter
(FD accumulator). The accumulator contents and status flag bits are sampled by
loading a buffer register upon receipt of a sample command. The buffer register
data is transmitted in eight bit groups (characters). The lower order eight
bits are loaded into the UART; after transmission, shift pulses load the next
character in the lower order location. The sequence is repeated until eight
characters (sixty-four bits) have been transmitted.
RBD-2
This card contains the parallel/serial converter (UART) and a command decoder.
Transfer of data from the buffer register to the modem is accomplished using a
'UART1 chip; UART is the abbreviation for 'Universal Asynchronous Receiver-
Transmitter*. The UART transmitter section converts parallel data into a serial
word consisting of eight bits of data in addition to start, parity, and stop bits.
Serial data from the UART is the modem SEND DATA input.
-------
-49-
I
1-3
w
H
CO
I
hJ
3
W
o
w
> P CM
25 p
5 M
0 P Crf
Q
Pd
w
PQ
U
H P
O S
!2!
o
M
r i v^t
M
-------
-50-
The UART receiver section converts serial data from the modem RECEIVED DATA
output into parallel data. Code transmission is verified by checking for data
parity and receipt of a valid stop bit. One eight bit parallel word is produced
for each character transmitted.
Command site transmissions are one character in length. The four possible commands
are encoded with redundancy in that two identical four bit bytes are transmitted.
The command decoder compares the received bytes at the UART receiver data output;
a command output is produced only if the received bytes are identical and received
data parity exists. These decoding requirements minimize the possibility of
initiating a false command at the receiver site. The command decoder outputs
source the "Acquire Reference1, 'Zero Accumulator1, and 'Sample1 pulses required
by the receiver site.
RBD-3
This receiver board generates the timing and control signals required by the
Data Acquisition Section. Timing signals required for modem operation are derived
from a crystal oscillator.
Modem
Data is exchanged with the command site using modems operating over leased telephone
lines; either two wire or four wire lines can be used by the proper modem
connections. The data rate used is 1200 baud, which provides the resolution re-
quired for the sample command and also permits future expansion of the data
message to include data from several receivers. The modems encode serial data
for transmission and decode received signals to serial data.
-------
-51-
D. Receiver Checkout Unit
A receiver checkout unit was designed to facilitate the checkout and maintenance
of the METRAC receiver. The unit is connected to the METRAC receiver by a multi-
conductor cable such that the contents of a receiver sample are displayed.
Receiver FD accumulator contents, station ID, and receiver FD , as well as
receiver status bits, are presently displayed. Provision has been made to in-
corporate displays for sensor data.
A photograph of the receiver checkout unit appears as Figure 20. The different
displays provided are located as follows. The top row provides seven digits of
display of FD accumulator contents, while the second row contains one hexadecimal
display for station ID. This row also contains a provision for two three-digit
groups of sensor data. The bottom row of display indicates FD , the rate of
change of the FD accumulator data. Two columns of four light emitting diodes
provide display of the receiver status bits. All displayed data are loaded on
receipt of a sample pulse from the receiver.
A block diagram of the checkout unit appears as Figure 21. Two circuit cards
are required. The card labeled MBDl contains the display devices, while MBD2
contains circuitry to load the receiver sample into the proper display, to
control the FD display, and to alternate display blanking (included to reduce
power requirements). The circuit cards and a power supply are mounted in an
instrument case. The unit operates with 117 volt AC power.
E. Command Site Console
The command site console controls the operation of the METRAC receiver site
equipment. The receiver network data (receiver samples) are sampled, then
buffered and formatted to produce a sample record which is recorded on a cassette
-------
u.
i
X 0
O M «
H W H
SUM
i
o
CM
a
Pn
-------
-53-
H ffi
u o
W H
hJ H
W E£
CO CO
i
,-4
§
H
53
O
O
PL,
CO
M
Q
O
H
T
[ FD ACCUMULATOR ^ ^ |
)
}
i-i
CM
CO
*
Ln
o
i
O
o
Pd
o
CO
^
w
CO
o o
o o
Q
M
1 1
CM
CO
,1
CM
^
1 FD DISPLAY
^-i
CM
en
*
m
o
-
[ MONITOR DISPLAY
CM
O
H
O
W
O
O
-------
-54-
tape data recorder. A data monitor function provides command console display of
selected receiver site data. The command console also contains a reference
transmitter control unit which provides remote control of the reference trans-
mitter site equipment. Figure 22 shows a block diagram of the equipment present
at the command site.
The command site initiates both manual and automatic commands for receiver
control by generating four independent commands. The first (and automatic)
command is the sample command. Pulses generated by a sample pulse generator
initiate the transmission of a sample command to all receiver sites. The trans-
mission time is concurrent for all channels. In response to this command
receiver sites will sample their data and automatically transmit this data to
the command site.
Three manual commands may be programmed on the command console panel. A
selected manual command may be transmitted to a selected receiver site or to
all receiver sites. To minimize the possibility of accidental command transmission,
a step sequence must be completed to send a command. First, the command is
selected with a two position toggle switch, after which a station select push
button is depressed. The command transmission occurs upon depression of a "Send
Command" push-button switch.
The three manual commands are "Zero accumulator", "acquire reference", and at
present, a spare . The zero accumulator command sets the FD accumulator at
selected receiver sites to a count of zero which facilitates the determination
of proper receiver operation previous to balloon release and provides a starting
point for position computation.
-------
-55-
hJ
o
PL,
3 W
la H
13 M
M CO
2
O
H
n
\
\
W W
H H
H
CO
H
y
o
w
PM
^-.^
§CO En
CJ
A
%
k
Q C
1
w w
H H
!=J &
P , A 1
CJ CJ
O
H
W
j
O
CJ ^
4
f
1
w
H
M
CO
I J
1
H
CO
M
a-
<
i
(
k ^
r 1
5J
:t!
-1
<
3
k J
r i
CJ
H
§
k ^
r \
k
2S
^
o
r
^
^
4
V
w
PM
1
d
i
s
CJ
^
A
^%
Jgrj
CJ
!/
<
f
P.
(/
I
e
h-
<
h
^>
(/
H-
>
)
ri
< Q Q
| M l-l td
I CO -
H
^
PC
1 -f, f£,
5 t3 E
c^
> M
00
5 W X
cj cj an H
co cj o H a
»
1 <£ <
: M w
> H fa a
i <
E-
3 o
H o e
CO PCI PC
4
w'
O
w
3
g
CO
w
1
M
k
M CJ
H W
co co
L '
k -
g
g g
a ^
S "^1
O H
^3
Q
j
g
H
M
!a
0
a
k
1
^
o
CJ
o w w PM
5 a 3 M
H 0 &
aa o o-
EH w
k M
" ft H
CO
( ( 1
k
^
X
W hJ
H w q
H CJ Prf
co S H
W S
§ W CJ
CJ fa
M
PH
O
CJ
0
$
o
&4
8
h4
pq
CM
W
^
O
h 1
S H^
^5 O
2 EH
w S
£t| Q
§ o
CO W
TV] QJ [V]
g O W H
H H 2 M
i-4 00
-------
-56-
The "acquire reference" command enables the search by receivers for the reference
signal. The presently unused spare command could be utilized in a future system
to control receiver power, etc.
Acquisition of data from receiver sites is initiated by the transmission of a
sample command. Receivers decoding this command will sample this data, buffer
it, and transmit it to the command site. Modems at the command site translate
this data into a serial stream. The serial asynchronous data are reformatted
into eight bit parallel words and stored in buffer registers. After all receivers
have responded, the data stored in the buffers is read out sequentially and
recorded by a cassette tape unit. One sample record of data from all receivers
is recorded for each sample pulse issued by the command site.
A sample pulse generator supplies the command site timing pulse which generates
the sample commands; the time interval between sample pulses is internally
programmable by means of a card mounted switch. The interval may be varied in
one second increments from a minimum of one second to a maximum of fifteen
seconds. This could permit a reduction in the amount of data generated during
a balloon flight. The selected time interval is produced by digitally dividing
the output of a crystal oscillator, which eliminates significant trajectory error
because time uncertainty is maintained less than + .017o.
The command site data monitor performs several system functions. First, this
function aids in the checkout and maintenance of a large segment of the METRAC
system as receiver and data acquisition equipment performance can be monitored
and the receiver sample for any receiver site may be displayed. In addition,
the data monitor provides reference transmitter frequency control information
during a balloon flight. A display of FD , the rate of change of the FD accumulator,
-------
-57-
indicates what reference transmitter frequency change is required to maintain a
nominal frequency difference. Finally, receiver site performance during a flight
may be monitored by means of a continuous display of the status bits for all
system receiver sites. Additionally, the difference in FD accumulator between
two receiver sites may be displayed, which permits determination of proper opera-
tion at all receivers at launch.
The reference transmitter control unit permits an operator at the command site
to turn transmitter carrier on or off for a balloon flight, testing, etc. The
unit also enables tuning the reference transmitter frequency to maintain the
nominal difference required for system operation. Information for frequency
tuning may be obtained by observing the FD rate associated with any receiver
site. The unit was designed to permit flexibility in that automatic and manual
frequency control is provided. The automatic system contains limited safeguards
to enhance operating reliability.
The command site console was designed mechanically to provide complete access to
the electronic hardware and interconnect. The logic circuitry associated with
command site data acquisition and the data monitor functions is mounted on a
swing out logic chassis and logic cards are plugged into connectors mounted in
this chassis. The various displays are mounted on the front panel which also
swings out for access to the circuitry. Three photographs illustrate the construct-
ion used.
A photograph of the command site console front panel (Figure 23) shows most of
the display units and control switches. The station data monitor contains the
dual three row displays in the upper section of the front panel. Switches to
control these displays are located directly below the displays. The status monitor
display is located below the station data monitor display. The command select
switches are at the left side of the panel, while station select switches are
located across the front panel below the status display. The reference transmitter
control unit was not installed at the time the photograph was taken.
-------
FIGURE 23. PHOTOGRAPH OF
COMMAND SITE
FRONT PANEL
-------
-59-
Figure 24 shows the construction of the front panel. The station data monitor
displays are located on two printed circuit cards, while the status monitor
display is located on a third printed circuit card. Flexible woven cable is
used to interconnect with the logic circuitry.
The rear view (Figure 25) shows the logic rack in an open position. The logic
circuitry swings open for checkout and maintenance. Modems are mounted below
this rack and the telephone line connections are made on the panel at the lower
left of the view. The connector for data output to the cassette tape is at the
top of the cabinet.
1. Data Acquisition Logic
The command site data acquisition logic accepts receiver network command inputs
from command console switches and the sample pulse generator for the transmission
to the receiver sites. To enhance transmission reliability, the commands are
encoded and transmitted to the receiver sites. Commands are encoded as two
identical four bit bytes and the resultant eight bit word is converted by a
UART from parallel to serial form. The serial character which includes data,
start, parity, and stop bits is the input to a modem which modulates a carrier
to transmit data over leased telephone lines to the receiver sites.
Receiver sites decoding the transmitted sample command transmit their data
message (receiver sample) to the command site and modems demodulate these trans-
missions forming a serial data stream. A UART performs a serial to parallel
conversion. The data are stored in a buffer register, sixty-four bits (8 char-
acters) in each channel buffer.
-------
FIGURE 24. PHOTOGRAPH OF COMMAND
SITE EQUIPMENT FRONT
PANEL CONSTRUCTION
-------
FIGURE 25. PHOTOGRAPH OF
COMMAND SITE
EQUIPMENT REAR VIEW
-------
-62-
With data stored in the command site buffers, the recording process can begin.
The outputs of the buffers are connected to a bus which is the data input to the
cassette tape recorder. The buffers are enabled in sequence to form a sequential
tape record. The buffer for Receiver #1 is enabled first and the cassette unit
incrementally records the eight (8 bit) characters contained in the buffer. The
process continues until data contained in each buffer has been recorded on the
cassette tape. A sync word, consisting of eight bits of data (all logic ones),
is written after the last buffer data entry to indicate the end of a sample record.
The cassette unit now waits until the next receiver sample is returned and the
data transferred signal is given. A new sample record is written by writing the
contents of each buffer. Upon termination of a flight record, the tape is
formatted into a 6600 compatible tape for processing.
Figure 26 shows the block diagram of the command site data acquisition logic.
Each receiver site in the network requires three circuit cards, consisting of a
modem, a modem interface card, and a register card. Two timing and control
cards generate the control signals required by the data acquisition logic section.
The modem interface card provides the interface required between the modem and
data acquisition logic and also contains the logic to produce the command signal
transmitted to receiver sites. The card contains the serial to parallel converter
indicated on the block diagram. A UART (Universal Asynchronous Transmitter
Receiver) converts the eight bit parallel data format used within the data
acquisition logic to/from the asynchronous serial data format at the modem.
Commands generated by the sample pulse-generator or command console switches are
encoded in the command encoder logic, producing an eight bit character for trans-
mission. The command is converted to a serial format and transmitted by the
modem upon receipt of a pulse from the sample pulse generator.
-------
w
H
C/3
hJ
^
q
V-'
'1
.
i-4
PM
A
- \
» ^
q
PH
\
ft
>
O
o
Pi
w
CO
oj
1
PM
j .
o-;
w
Cu
J .
>
0
>- >
&i
PM
W
CO
t
!v;
^
t
* * *
f
I
1
^
H P
CASSE
t
w
fr
fe
f
t
fc
o
o
Pi
(i
"p?
w
CO
(
MODEM
M
i-4 CO
J
W Pn
^? ta-4
o
ffi O
PM H
q
LI
-^
-63-
§
M
s
W
r
o
M
E-i
M
CO
M
O1
tff U
£
^
W
E-i
H
s
O
c
JJ
^ o
T! rT _,
H9
^ 3 , ,
1 "?
1 g c,
i j PI i"-i
S 3
CO J
H CM
<3 PM
Q O
P4 O
f
T
PM
CO
M
STATION
MONITOR E
|
T
>H 1-4
3 9
hJ Pi
PM H
co g:
M O
P U
>' ^
u
Pi
o
co
CO
W
P U
b q
pi
PH
r
H
CJ
9^
w q
CO Pi
Pi iZ
> o
o o
pi
Pi
O
E-i
H
1
.J
0
PS
g
o
o
H Q
b
*\
§ Q
5 w
0 H
0 C3
< PM
o o
P"4 O
*
T
PM
CO
M
STATION
MONITOR I
i
T
>< *4
0
u o
erf
CO
SWITC
a
2
§
H
H
M
tfl
>H
C3
Q»
3
^
f*
<
P
H
H
H
CJ
IAGRAM
6
V
8
, *)
pq
v _>
w I
S
t>
O
M
fe
Pi W
> H
OHM
Pi co
-------
-64-
When receiving data from the receiver sites, the modem data output is converted
from serial to parallel by the UART. The modem interface card controls the
loading of this data into the register card. The eight parallel UART data out-
put lines are the inputs to the register card, such that the register is loaded
with eight bits (one character) each time the UART indicates a data character is
ready (DR pulse). A character counter is advanced by the data ready pulses;
when eight characters have been counted, the transition to a count of eight fires
a one shot to generate a "Data Received #1' pulse. The command site control
logic uses this pulse to determine when data is buffered and ready for sequential
output. Control signals enable the transfer of data from the buffer and data is
shifted out sequentially by character.
The register card contains the 64-bit buffer which stores the received sample
from the receiver sites. The buffer is constructed as a shift register eight
stages in length with eight bits of parallel storage per stage. The buffer
input is eight bit parallel data from the UART, while the buffer output is to
the bus providing eight bit parallel data to the tape unit and command site data
monitor. Control signals from the monitor interface card control the register.
The bus logic card (one of the two timing and control cards) controls the readout
of data from the buffer registers and also supplies timing pulses required for
UART operation. 'Data Received #l' pulses from all monitor interface cards are
wire or'd together. The first such pulse sets the tape unit in a write mode and
clears a counter (Receiver Enable Counter). Data from receiver #1 is read out
of the buffer and written on tape. The counter is then advanced and data from
receiver #2 is written. The process continues until data from all receivers is
written on tape.
-------
-65-
Cassette Tapes
The cassette recorders used in METRAC are incremental digital cassette storage
and retrieval systems which record and playback eight bit characters. When
writing, the recorder accepts a parallel eight bit character and a start signal
at TTL logic levels. When reading, each start signal will cause one eight bit
character to be read and presented in parallel at the output, together with a
strobe signal. Outputs are TTL levels.
Data is recorded on standard Phillips certified tape cassettes in complementary
NRZI format. The recorder operates incrementally at rates up to 120 characters
per second when recording and up to 100 characters per second when reading. When
read continuously, as is done in METRAC, the output data rate is approximately
110 cps.
When writing, the first character on tape is Cl (Character 1) of receiver #1.
Then the remaining characters of receiver #1 are written, followed by the
characters of receiver #2, and so on. After a complete Sample Record is written,
consisting of 56 characters (8 Char/receiver x 7 receivers), a sync word is
written consisting of all ones. This gives a total of 57 (8-bit) characters per
Sample Record.
2. Data Monitor
The command site data monitor, located within the command console, contains
three display units. The status monitor display continuously indicates the con-
dition of the eight status bits associated with each receiver site for up to 16
sites.
-------
-66-
Two identical and independent station data monitors each provide display of a
receiver sample (FD accumulator, ID, provision for sensor data) and FD , the
difference between FD accumulator samples. Each monitor contains two one-of-
sixteen station select switches which are labeled W and X, Y and Z. A third
switch enables display of the receiver sample for: 1) Station Select W (or Y);
2) Station Select X (or Z); and 3) Display Alternates between W and X (Y and
Z). The FD display for 1) and 2) indicates FD ,' the change in FD accumulator
between two consecutive samples and in 3), the FD display indicates the
difference in FD accumulator data between station W and station X or station Y
o
and station Z.
The command site data monitor block diagram appears in Figure 27. Here signals
from the Data Acquisition Section (data bus and timing) pass through the monitor
interface to the data monitor cards. The status display card latches the eight
bits of status information in each receiver sample and light emitting diodes
provide the visual display.
Each station data monitor contains four cards. The station monitor display card
contains an upper row of seven latching BCD readouts for FD accumulator data,
a center row with a latching hexadecimal ID display and space for two groups of
three latching BCD readouts for sensor data, and a lower row of seven counter/
latch BCD readouts for FD displays. The receiver select control card compares
station select switches with timing pulses; a pulse, MR, enables the display control
card to load the receiver sample into the monitor display card. The receiver select
card also controls the operation of the FD processor card. FD accumulator data
are loaded and the difference (FD ) is computed and displayed by the monitor dis-
play card.
-------
-67-
NUAL
SW
CM
O
-------
-68-
3. Reference Transmitter Control
The reference transmitter control unit is housed within the command site console
to provide carrier on/off and frequency control. The carrier is turned on or
off with a front panel switch and indicator lights show the carrier condition.
The reference transmitter frequency may be controlled manually or automatically.
In the manual mode of operation the operator observes the FD display for a
selected station and adjusts a ten-turn potentiometer to maintain a nominal
frequency offset. In the automatic mode, an AFC loop is formed by the
reference transmitter, a selected receiver site, and the command console. The
operator selects the receiver site to be used for frequency control by operating
the station select switches in the W(X) station data monitor; a primary site
and back-up site are selected.
A front panel tuning meter indicates the difference in control voltage between
the manual control potentiometer and the AFC circuitry. The operator nulls this
meter so that a changeover to manual control can be made without disrupting
system operation. Mode of operation is selected by a front panel switch and
L.E.D.'s indicate the mode in use. Operation will transfer to the manual mode
if five seconds elapse with FD outside the range of 2 to 8 KHz.
Figure 28 shows the hardware to realize the control function consisting of the
mentioned controls, indicators and meters, a frequency control card, and a
remote control unit. The frequency control card accepts inputs for frequency
control from the manual control pot and from the FD display of the ₯(X) station
data monitor. The AFC path uses a D/A converter and error amplifier to provide
the frequency control voltage to the remote control unit. The frequency control
voltage from the control card and the carrier control commands are encoded for
voice grade telephone line transmission to the reference transmitter site using
components manufactured by CDC Autocon. The control voltage is the input to a
-------
-69-
OO
CM
o
-------
-70-
voltage to frequency converter such that the varying frequency output keys a
narrow-band FSK tone channel transmitter. The transmitter outputs are combined
and connected to the telephone line.
F. Minneapolis Site
The Minneapolis site is the final stage in the process to convert receiver site
data to a computer compatible format. Data produced by the command site and
written on magnetic tape cassettes is translated into CDC 6600 compatible tapes.
In order to expedite the conversion process, a data retrieval link capability is
included in the Minneapolis site equipment. The cassette data recorder at the
command site, a telephone line, and the Minneapolis site equipment allow the
duplication of the command site tape.
Figure 29 shows the block diagram of the data link. At the command site a completed
cassette is loaded and read by the cassette recording unit. This data is in an
eight bit parallel format, with a UART used to produce a serial data format. A
1200 baud modem then encodes the data for transmission. Data into the Minneapolis
site is demodulated by an identical modem, formatted by a UART into parallel
data, and recorded on another cassette tape unit. The data is transferred between
tapes at 1200 baud.
Tape cassettes produced by the command site and transferred to the Minneapolis
site either by the data link or by physically transporting the cassette are then
reformatted. Figure 30 illustrates the functions in the process. The eight bit
characters on the cassette are converted into six bit words for 6600 use. The
data are reformatted by reading three 8 bit characters from the cassette into
a 24-bit buffer. The data are then read out as four 6-bit words and written on
a Kennedy 1708 tape unit, producing the desired tape.
-------
-71-
CAS-
SETTE
RE-
CORD-
ER
PARALLEL
DATA
PAR/
SER
CONV.
SERIAL
DATA
MODEM
TIMING
LOGIC
TELEPHONE
LINE
DATA LINK FROM COMMAND SITE TO PHONE LINE
TELEPHONE
LINE
MODEM
*.
*.
CARRIER "
DETECTED
r
SE
P
CO
I
R/
AR
NV.
k
k.
TIMING
LOGIC
»
CAS-
SETTE
RE-
CORD-
ER
i
i
PHONE LINE TO MINNEAPOLIS SITE DATA LINK
FIGURE 29: BLOCK DIAGRAM DATA LINK
-------
-72-
oo
H
M
W
fd
H Pi P4
H O O
Pi W W
H PS!
M W
CM FQ
CO
W
I
PH
§
H
H
W
o
O
O
vO
VO
-------
-73-
As the data is read from the cassette, the buffer input looks for a valid sync
word consisting of all ones (one character of all logic ones). The first sync
detected will enable the timing circuit to gate in the first three characters
(after sync) to the buffer. The six parallel bit words are clocked out
sequentially at a 4.8 KHz rate and recorded on the Kennedy 1708 tape transport;
this rate enables the cassette recorder to read continuously to minimize tape
production time.
Reading of the cassette and recording by the Kennedy transport continues until
all data is transferred to the Kennedy. Sensing that data is no longer coming
from tape, the interface enables an end of file to be written on the tape. The
tape made on the Kennedy is then ready for processing by a 6600 computer.
TM
VI. METRAC POSITION DETERMINATION
A. Discussion of the Problem
The problem of computing the location of a balloon borne transmitter by measuring
accumulated differential doppler frequency between pairs of fixed land based
receiver stations reduces very naturally to a problem of solving a set of hyper-
bolic equations. This type of solution is common in such navigational techniques
as LORAN and OMEGA.
To describe the METRAC principle, consider the two dimensional example depicted
in Figure 31. This figure shows a sample set of hyperbolas formed with two
receivers serving as the foci. As a transmitter is moved from Point A to Point B,
receiver 2 accumulates a frequency count larger than the transmitted number of
cycles while receiver 1 accumulates a smaller count. If the transmitter were to
move along one of the hyperbolas, both receivers 1 and 2 will count the exact
-------
-74-
RCVR*
RCVR*2
FIGURE 31. Hyperbolas Formed By Two Receivers
-------
-75-
number of cycles transmitted. The difference in the number of cycles counted
between two receivers such as 1 and 2 is called the differential doppler and is
equal to twice the number of hyperbolas crossed by the transmitter.
Since the number of cycles counted at the receivers is an integer number, there
are exactly dMA hyperbolas defined by two receivers separated by a distance d
and a transmitter operating at a wavelength A. and a frequency multiplication
factor of M. The separation of the hyperbolas on the axis joining the two re-
ceivers is A./M. For a frequency of 403 MHz and a multiplication factor of 8 as
was utilized in this work, the separation is about 10 cm. This corresponds to
the finest position resolution possible for the given frequency and multiplica-
tion factor.
The differential doppler accumulated over a fixed time interval is independent
of the path taken. In addition, there are a doubly infinite number of possible
starting and stopping positions A and B which yield the same differential doppler.
By knowing the starting point and obtaining two independent measures of the
differential doppler (i.e., by using three receiver stations), one can compute
the final position in terms of the intersection of appropriate sets of hyperbolas
as is shown in Figure 32.
For practical balloon tracking, we clearly need to be able to specify a three
dimensional position. The preceding discussions easily extend to three dimensions
for which case the hyperbolas become hyperbolas of revolution.
Four stations are then required to define three measurements of differential
doppler from which the solution can be derived. Mathematically, there are two
valid solutions to this problem. One, however, is generally underground.
-------
-76-
RC\/R*£
BALLOON
POSITION
RC'/R
FIGURE 32. Intersecting Hyperbolas Formed by Two
Receiver Pairs
-------
-77-
Using more than four receiver stations adds redundancy so that the momentary
loss of a receiver doesn't effectively terminate a flight. Whereas four
receivers yield one physical solution, five receivers yield five, six receivers
yield fifteen and seven receivers yield thirty-five solutions.
B. Effect of Geometry
As can be seen in Figures 31 and 32, the spacing between hyperbolic shells depends
upon geometrical position. Only on the line between each pair of receivers is
the separation of the shells as small as A./M. Clearly the accuracy of the com-
puted position depends upon the spacing of the shells as well as the orientation
of the foci (receivers) with respect to each other and to the position of the
transmitter. Maximum practical resolution inside the receiver array is achieved
by deploying an equally spaced ring of receivers about a central receiver. A
minimum receiver array should consist of a triangle of receivers with one in
the center. The centrally located station is very important in giving good verti-
cal resolution. When there are more than a minimum set of four receivers operating
properly over some time interval, some of the solutions will be better than others
because of the differences in the geometry.
Well outside the borders of any receiver array, the hyperbolic shells tend to
become nearly parallel to one another. This implies that the uncertainty in
computing position of the transmitter grows as the balloon drifts away from the
array. The uncertainty is largest in the radial direction. Increasing the
averaging times in determining the winds outside of the station array will help
to decrease the effect of the position uncertainty.
-------
-78-
C. Sources of Error
There are two primary sources of error which can affect the METRAC solution:
count errors and station location errors. Count errors can come from a variety
of sources, some interrelated. The most serious errors would arise if either
the balloon or reference transmitter were interrupted even momentarily. This
would cause dropped counts at all receiver stations simultaneously so that no
solution could be computed.
Count errors will also occur if the frequency difference (reference and balloon
transmitter) tracking filter loses lock at any receiver. This will happen when-
ever the signal to noise ratio at the receiver becomes sufficiently small for an
appreciable part of the time constant of the filter. Experience during the
Minneapolis field tests showed that this situation occurred most often when the
transmitter was high and nearly directly above a receiver station or when the trans-
mitter was far (30-100km) from the receiver. If the signal to noise ratio becomes
small for a very short period of time, errors may occur in the counting even
though the filter track remains locked to the frequency difference. Tests per-
formed with a static transmitter showed that this type of error was generally
random with magnitude of only one or two counts. Errors in FD counts can also
occur due to sampling uncertainty. However, these errors are non-accumulative
and are also on the order of only one or two counts per sample.
An error from any of these sources will locate the balloon transmitter between
the wrong hyperbolic shells. If the error is not random, as is the case for the
first two errors described above, the position error will grow as the balloon
gets further away from the array and the hyperbolic shells get further apart.
Random errors due to sampling uncertainty or occasional count errors will add
artificial variance into the true position, but unless the balloon is well outside
-------
-79-
the baseline where the shells are far apart, this error in velocity is small when
even computed over very short time intervals.
In addition to position errors due to inaccuracies in the count of the multiplied
offset frequency difference, errors also arise from uncertainties in the locations
of the receiver stations themselves. The solution to the METRAC positioning
problem requires frequency counts at a minimum of four receivers as well as
initial launch coordinates relative to the receiver array. Because of the extreme
resolution inherent to the tracking system, an error of ten meters in the
location of one station relative to the rest can be equivalent to as much as
a 100 count sampling error. As the balloon moves away from its starting location,
the position error will grow because the hyperbolic shells become more widely
spaced as was discussed above.
Figures 33 and 34 present the results of a computer simulation of position errors
caused by the mislocation of receivers. Figure 33 shows a triangular array of
receivers with an X-Y plot of the trajectory of a balloon launched from the
central receiver station. The balloon is carried outside of the receiver array
at a height of about 2200 meters. Figure 34 presents three examples of how the
computed height deviates from the true height of the transmitter after three sets
of rectangularly distributed random errors ranging from -1 to +1 meters have been
added to each coordinate of each receiver station. An error in the launch location
with respect to a perfectly specified set of receiver stations causes similar
errors.
D. METRAC Solution
Several techniques have been utilized to solve the basic hyperbolic navigational
equations. Relatively straight forward iterative solutions are discussed by
Ranter (1962) and Newton (1967). Newton (1967), Razin (1967) and Glish (1971)
describe various closed-form solutions with varying degrees of complexity and
-------
-80-
1km
km
FIGURE 33. SAMPLE BALLOON TRAJECTORY FOR
TRIANGULAR RECEIVER ARRAY
-------
-81-
4000
Case
3000
I
o
LU
2000
1000
0
1
Case 3
I
-20
-10
HEIGHT
0 10
ERROR (m)
20
FIGURE 34. THREE EXAMPLES OF HEIGHT ERROR VERSUS
HEIGHT FOR RECEIVER POSITION INACCURACY.
RANDOM ERRORS TAKEN FROM A RECTANGULAR
DISTRIBUTION (-1, 1) WERE ADDED TO EACH
KNOWN RECEIVER LOCATION.
-------
-82-
accuracy. More recently a novel approach has been developed by Schmidt (1972),
and it was this method which was used to compute the METRAC solutions.
The classical formulation of the navigational problem is to determine a position
when the range difference to two known stations is measured. Schmidt showed that
the problem could be alternatively formulated in that the desired solution is
the focus of a general conic whose major axis is defined in terms of the
differences in range to three known stations. This method is particularly
attractive not only because it is a closed-form solution, but also because it
reduces the basic problem to solving a set of simultaneous linear equations.
E. Software
1. Software Input
The recorded data output from the METRAC command site consist of one second
values of accumulated (multiplied and offset) frequency difference counts
between the balloon-borne transmitter and the reference transmitter measured at
each receiver site, FD , as well as status information from each receiver. The
numbers of relevance in computing the balloon position are the accumulated
Doppler counts, a small part of FD . The bulk of FD is due to the normal quasi-
constant difference in frequency between the balloon and reference transmitters
and is common to all receivers during each sample. By forming the differences of
FD between pairs of receivers, this large component cancels out leaving the
desired Doppler difference counts. All status information is also used to
evaluate the quality of each data sample.
-------
-83-
Status consists of the following:
1. Reference signal acquired
2. AFC
3. Filter track
4. F, < F, MIN
d d
5. F , > F , MAX
d d
The first three status bits are normally 1 and the latter two are normally 0.
2. Program METRAC
Figure 35 shows a general flow diagram of the METRAC software package. The
following sections briefly describe each of the blocks of this diagram.
METRAC
Program METRAC is the primary controller to the program flow. During the first
pass, initialization and definition routines are called. Subsequently METRAC
cycles until either a flight duration limit (ITIME) or an end-of-file is encountered.
READIN
Subroutine READIN inputs the following initial data:
(1) Flight identification - 80 characters of free field alphanumeric
characters.
(2) Number of receiver stations used.
(3) Serial number and relative x, y, z location of each receiver station.
(4) Transmitter frequency (403 MHz).
(5) Frequency multiplication factor (8).
(6) Flight duration limit (seconds).
-------
-84-
ITIME = ITIME + 1
FIGURE 35: FLOW DIAGRAM OF METRAC SOFTWARE PACKAGE
-------
-85-
INITIAL
Subroutine INITIAL computes all the initialization and construct parameters used
throughout the remainder of the computations. These include:
(1) Effective wavelength (wavelength/multiplication factor).
(2) Distances between each pair of receiver stations.
(3) Distances between launch point and receiver stations.
(4) Initial difference frequency accumulator values.
UNCODE
Subroutine UNCODE reads a record (one second sample) of data from magnetic tape
and decodes it to yield the accumulator and status data.
STATUS
Subroutine STATUS determines which receivers have had status bits during the
sample so that they will not be used in computing the transmitter position.
ACCADJ
Subroutine ACCADJ differences the new accumulator values from the previous value
and adds them to a bank of working accumulators. The working accumulators are
structured such that the difference between any two of them is proportional to
the difference in distance from the receivers to the transmitter.
DOPDIF
Subroutine DOPDIF forms the Doppler differences between all pair of receivers.
-------
-86-
COMPUTE
Subroutine COMPUTE used the input station locations and the Doppler differences
to compute all possible METRAC transmitter positions. If all seven receivers
send valid data, 35 solutions are computed for each time sample.
DECIDE
Subroutine DECIDE uses a primitive scheme of maximizing Doppler differences in
choosing the "best" solution obtained from COMPUTE.
PATCH
Subroutine PATCH uses the solution obtained from DECIDE to compute a new
accumulator value when a receiver has had a status error. That receiver can
then again be used in subsequent samples.
POSWND and STATOUT
Subroutines POSWND and STATOUT print out all position, wind and status data
computed for each sample.
VII. MINNEAPOLIS FIELD TEST
This chapter presents the results of a field test of the METRAC System conducted
in Minneapolis by Control Data Corporation. The test was performed in order to
complete the evaluation of the feasibility of utilizing METRAC to obtain accurate
detailed measurements of the urban wind field. Since no radar data was available
for comparison with the METRAC data, the only independent verification of the
-------
-87-
system's accuracy was obtained by comparison with winds derived by tracking a
radiosonde package attached with the METRAC package to the same balloon. For
this comparison a WeatherMeasure RD-65 rawinsonde system was borrowed from the
University of Wisconsin.
TM
A. METRAC Deployment for the Minneapolis Test
Planning for the METRAC test began in December. It was decided to locate the
reference transmitter on top of the IDS tower and a lease was signed with
Chicago Broadcast Services. The reference antenna was installed after the first
of January and transmitter equipment was installed about a month later.
The following sites were selected as suitable locations for METRAC receivers:
(1) Control Data
Roof of South Tower
Minneapolis, Headquarters
(2) Radisson South
7800 Normandale Blvd.
Bloomington
South Penthouse Roof
(3) Prudential Insurance
North Central Home Office
3701 Wayzata Blvd.
Minneapolis
Tower roof
(4) Kenneth S. Gage (private home)
4833 Aldrich Avenue So.
Minneapolis
30 foot tower guyed to roof
-------
-88-
(5) Model Ready Mix Co.
400 W. 61st Street
Minneapolis
Top of radio tower
(6) General Communications - St. Paul Shop
756 N. Snelling
St. Paul
Top of radio tower
(7) Imperial Heights Apartments
90 Imperial Dr.
West St. Paul
Tripod on roof
Figure 36 shows the locations of the reference transmitter and the receivers on
a map of the Twin Cities. Each R on this figure represents the location of a
receiver and X represents the location of the reference transmitter. Installa-
tion of the receiving equipment was completed for six stations by the end of
February and the seventh station was added in the middle of March. A network
of leased telephone lines was installed by Northwestern Bell to connect each
receiver to the command site which was located in the Research Division's
METRAC laboratory.
A cassette tape recorder was set up at the command site to record raw data from
the station network. Equipment was also set up to enable the transfer of raw
data to a format which could be read directly into the CDC 6600. In order to
compute the location of the mobile transmitter the METRAC algorithm requires the
initial position of the transmitter and the relative locations of all receiver
sites. For the purposes of the Minneapolis test this information was obtained
from survey maps which are available from Municipal, County and State Survey
Offices. Of special value were the "100-scale" maps from which horizontal loca-
tion can be determined to within a few meters. These maps also contain elevation
-------
-89-
w
H
H
H
O
t
E
.*
o
^^»
\
m
r H
i 13
Pi
O
h
w
O
H
H
/ ^
» J
H
-------
-90-
contours every two feet. Coordinates for all locations were compiled in the
Minnesota State grid-south zone in order to provide a common frame of reference
for all stations. No special survey was attempted for the METKAC receiver
locations and therefore the uncertainty to which these locations are known remains
a few meters. Since systematic errors are known to result from location errors
of station position, an accurate survey is important for system calibration.
B. Preliminary System Tests
The METRAC System was deployed for the field test late in February. Less than
two weeks were required to install receivers at six sites, to install the
reference transmitter, and to check reception at each receiver site. The latter
was accomplished by comparing the strength of signals received from the
reference transmitter and from a transmitter installed at the balloon launch
site, (on top of the Radisson South Hotel) with the strength computed for free-
space propagation.
As soon as several receivers had been installed, a system performance was checked
by acquiring the reference signal at each receiver and locking on to the second
transmitter located on top of the hotel. The commands to acquire reference and
lock were given from the command site. Our first indication that the system was
working properly came from the observation that the difference frequencies be-
tween each pair of receivers were nearly identical.
Satisfied with the results of the static test, our next step was to record data
as the METRAC transmitter was moved along a prescribed path on top of the pent-
hour roof of the hotel. The results of this walk are shown in Figure 37 where
every one second data point is plotted. There seems to be a systematic discrepancy
of about a meter between the path walked and the path computer by METRAC. Such an
"error" may be due to the fact the transmitter was secured at the end of a,long
-------
-91-
C\l
C\J
CD
O
(UJ)
C\J
I
co
O
-------
-92-
hand-held pole which is difficult to hold exactly vertical. Another possible
explanation is the position location errors mentioned earlier and the need for
system "Calibration" by more accurately surveying receiving antenna locations.
The data that were recorded from these preliminary system tests provided the
first realistic test of the data processing system.
C. Wind Profile Comparison Tests
A test was carried out in the Minneapolis area during April 1974 to compare wind
profiles obtained from the METRAC system with wind profiles obtained from
rawinsonde and theodolite measurements. During this period, eight balloons were
launched from the top of a 22 story suburban hotel. Each balloon carried both
the lightweight METRAC transmitter and a standard 1680 MHz VIZ radiosonde. The
radiosonde package was tracked with a portable WeatherMeasure RD-65 rawinsonde
system on loan from the University of Wisconsin Department of Meteorology,
Madison, Wisconsin. In addition, a theodolite was used to track the balloon
optically when cloud cover and visibility permitted. This section presents typical
METRAC data and wind profile comparison data obtained in the Minneapolis test.
Figure 38 shows the x-y trajectories for the comparison data presented in this
section. The locations of the seven receiver stations are also shown. The
trajectories labeled MF3, MF4 and MF7 represent six minutes of data, and traject-
ories labeled MF2 and MF5 represent twenty-eight minutes of data. Both flights
MF2 and MF5 extend well outside the receiver station array.
Figures 39 through 44 show the wind profiles computed for trajectories MF3, MF4
and MF7. Each of these figures presents the comparisons between METRAC derived
-------
-93-
I
10
H
6
CO
o
w
en
Pi
o
CO
H
pi
o
H
CJ
w
H
I
00
CO
C1)
i-l
-------
-94-
O
o
If)
(01)
o
o
o
1H9I3H
O
O
if)
ro
LL
I
I
T
u u u
o
CM
If)
-------
-95-
(LU) 1H9I3H
O
O
O
eg
O
O
o
o
o
o
o
ID
U U U
-------
-96-
O
O
m
1H9I3H
o
O
o
O
O
in
I
I
u u u
-------
LL
-97-
H9 I3H
o
o
O
O
O
o
o
in
UJ
i-
I
O
O
m
iO
o
u
^
W Cn
CO
cri
o o
PH CO
O [-3
O fe
CO O
O M
O S
(N
o
M
fc
-------
-98-
O
O
O
CM
O O
CM CM
UJ
O
O
in
1 H9I3H
O
O
O
o
o
in
if)
U
-------
-99-
o
o
If)
O
O
O
H9I3H
O
O
I
u u
Pi
O O
CM PH
fn w
O td
O fc;
C/3 O
M pi
Pi p*
Q
O
-------
-100-
wind profiles and rawinsonde and theodolite derived wind profiles. Figures 39,
40, and 41 show the first six minutes of flight observations plotted once per
minute, and Figures 42, 43 and 44 show the same flights with observations plotted
at twenty-second intervals. Since rawinsonde measurements were taken only once
per minute, they are not included in the latter figures.
Rawinsonde and theodolite winds are determined from measurements of elevation
and azimuth angles and independently computed or inferred values of height. The
accuracy to which a radiosonde system can determine these angles is dependent
upon the beam size of the antenna. Great precision generally requires the use
of large antennas and complicated pedestal machinery. The RD-65 rawinsonde
system has a minimum resolvable element of 0.1 and experience shows that the
RMS error may be as large as several tenths of a degree. Optical theodolite
tracking with an experienced observer is substantially more accurate with an
RMS error of only a few hundredths of a degree. For comparison, the RMS error
often associated with the GMD-1 rawinsonde system is 0.05 (Danielsen and
Duquet, 1967).
Uncertainties in the determination of the height of the balloon also affect the
accuracy of the horizontal position computed from azimuth and elevation angles.
This error is particularly significant at low elevation angles. For the Minnea-
polis tests, thermodynamic heights computed from the radiosonde data were used
for both the rawinsonde and theodolite determined winds. Errors in timing the
angular measurements also look exactly like height errors in the computation and
are critical when the angular position of the balloon is changing rapidly. This
problem is most serious in the early parts of the flight and with strong winds.
The errors described above can easily account for errors in the wind speeds of
1-3 m/sec for 60 second unaveraged rawinsonde winds. The errors will be largest
in the radial direction due to uncertainties in balloon height and will be
dependent on wind speed as described above. These facts may explain the largest
discrepancy between METRAC winds and the rawinsonde winds which occur in the u
component of Figure 39.
-------
-101-
Because radiosonde heights were also used in determining the theodolite winds,
the normal single theodolite pibal tracking assumption of a known constant
ascent rate was unnecessary. This assumption is particularly bad near the earth's
surface in an urban environment. Double theodolite techniques must be employed
to compute accurate pibal winds such as those presented by Ackerman (1974).
However, by using radiosonde heights, sixty-second theodolite wind errors should
be smaller than 1 m/s. Figures 43 and 44 show excellent agreement between
METRAC and theodolite winds even over twenty-second intervals except for
isolated values which were read or recorded incorrectly.
METRAC wind accuracy is limited in a way very different from the rawinsonde or
theodolite accuracy. The differential Doppler numbers locate the balloon borne
transmitter between two hyperbolic shells found by rotating hyperbolas about an
axis joining a pair of receivers (foci). The three-dimensional position can
then be solved as the common volume formed by the intersection of three such
regions. Clearly the best resolution in position is obtained when this volume
is smallest, and this occurs when the hyperbolic shells intersect one another
orthogonally. Position errors can be large when the intersecting shells are
nearly tangent. The minimum spacing between two shells for the Minneapolis tests
was 10 cm. When the transmitter is within the receiver array, the geometry of
the intersections is favorable. Simulation studies show that expected accuracy
for this case is within a few centimeters/sec for 10 second winds.
Actual errors in computed position and wind speed can occur if the Doppler cycles
are not counted accurately. This happens when the signal to noise ratio becomes
too small. In practice, this problem was uncommon and occurred only when the
balloon transmitter was well outside the receiver array or located almost directly
above a receiver. Utilizing more than the minimum number of four receivers cir-
cumvents this problem.
-------
-102-
Errors in computed position can also arise if the relative locations of the
receivers are not accurately known. This error is most serious for long flights
when the transmitter moves across the entire array. However, this problem can
be overcome by accurate station surveying.
Figure 39 thru 44 showed only the first six minutes of data or nearly two
kilometers in height. This is the region of most direct interest in air pollution
study. However, the METRAC system can track the transmitter to very high altitudes
and distances well outside of the baseline as was shown by trajectories MF2 and
MF5 in Figure 38. Figures 45 and 46 show the u and v wind components for these
two wind profiles compared with the associated rawinsonde wind profiles. The
agreement between the METRAC winds and the radiosonde winds is again excellent.
In fact, one can observe the apparent increasing amplitude oscillations in the
radiosonde winds in the top third of the wind profiles in both figures. These
oscillations are due to increasing errors in positioning the balloon due to the
fixed angular resolution of the rawinsonde system as well as the larger errors
associated with low elevation angles. Both errors are characteristic of single
dish tracking systems.
Errors in position computed from the METRAC system also degrade when the transmitter
is outside the receiver array. This effect, however, does not generally become
significant until one is several times the baseline of the receiver array outside
of the array and is probably undetectable in sixty-second or even twenty-second
wind profiles. For example, Figure 47 shows a comparison between twenty-second
METRAC winds and twenty-second theodolite winds between 800 and 1150 seconds into
flight MF5 (see also Figure 46). The comparison is excellent except for one bad
theodolite reading at 960 seconds into the flight. The largest discrepancy is
1.2 m/s which, indeed, speaks well for the theodolite observer. The balloon is
between 2 and 4 km outside of the receiver baseline and between 5 and 8 km from
the launch point and theodolite during this period of data.
-------
-103-
28
24
c
2°
UJ
16
1 2
8
I I MF 21
mm
METRAC (60sec)
RADIOSONDE (60sec)
I
I
1 2
1 0
8
0
LJ
6 I
10 20 30 -10 0 10
u (m / sec) v (m/sec)
FIGURE 45. WIND PROFILES FOR FLIGHT MF2
-------
-104-
I MF 5l
10
1 600
METRAC (60sec)
RADIOSONDE (60 sec)
8
1 200
UJ
800
400
6 x
0
LU
I
I
1
I
I
10 0 -10 -20 -10 0
u (m/sec) v (m/sec)
FIGURE 46. WIND PROFILES FOR FLIGHT MF5
-------
105-
O
o
m
CO
(W) 1 H 913 H
O
O
O
CD
O
O
m
in
o
o
o
m
T
T
T
m
u
-------
-106-
D. Special Capabilities of the METRAC System
Besides accuracy, one of the key features of the METRAC system alluded to above
is the capability of small scale resolution. Relatively small scale structure
can be exposed by decreasing the sampling interval. Figure 48 shows a rather
unexciting profile of wind (v component) which is nearly zero up to 3.0km as
measured by both the METRAC system and the rawinsonde system. Figure 49 compares
sixty-second METRAC determined winds with thirty-second METRAC winds, while
Figure 50 compares the sixty-second winds with fifteen-second METRAC winds. In
each of these figures, the presence of a strong and relatively sharp shear layer
emerges between 1 and 1.4km of height. Figure 51 illustrates this feature of
high resolution again with an example from flight MF7. The strong shear present
in the u component of the wind is shown to be concentrated in an extremely thin
layer of approximately fifty meters depth.
Figure 52 illustrates the optimal resolutions of the METRAC system when position
is computed for each one-second sample. This sixty-second segment of data with
"wind" speeds plotted every second shows the circular rotation of the METRAC
transmitter suspended below the balloon. The frequency and amplitude of this
periodic motion agrees favorably with simple pendulum theory. The detail of the
motion appears to be at least as good as the detail of the balloon induced
oscillations which have been measured with the FPS-16 Radar/Jimsphere system and
discussed by DeMandel and Krivo (1972). This kind of resolution encourages
further evaluation of the use of the METRAC system to obtain measurements of
atmospheric turbulence or of even looking more closely at the response dynamics
of the balloon as it ascends through the atmosphere.
Another important aspect of the METRAC system is that the solution is based purely
on geometry and the physical laws of electromagnetic wave propagation. In other
words, determination of the three balloon coordinates x, y and z are made without
-------
-107-
IMF 2
-METRAC (60sec)
-RADIOSONDE (60 sec)
LU
4
0
til
I
1
1
1
-10 -505
V (m/sec)
10
FIGURE 48. WIND PROFILE FOR FLIGHT MF2
-------
-108-
I MF 2
7
ME TR AC (60sec)
M E TR AC (30sec)
c
E
UJ
h-
i
CD
LU
I
I
1
-10
-5
V
0 5
(m/sec)
10
FIGURE 49. COMPARISON OF 60 SECOND AND
30 SECOND PROFILES
-------
-109-
C
E
LLJ
I MF 2
METRAC (60 sec)
METRAC (15sec)
1
-10 -505
V (m/sec)
10
t-
i
O
LJ
I
FIGURE 50. COMPARISON OF 15 SECOND AND 60 SECOND PROFILES
(MF2)
-------
-110-
o
o
o
CM
O
O
in
1H9I3H
O
O
O
O
O
m
I
T
I
I
o
o
o
o
C\J
o
o
m
u
0
^E
>
o
o
CM
U
0)
o £
co
H
s
o
o
w
-------
-111-
-------
-112-
any assumption of balloon ascent rate or hydrostatic equilibrium. This feature
combined with the high degree of resolution allows the possibility of measuring
the atmospheric vertical motion. As an example, Figure 53 shows the variation
in the vertical velocity of the balloon computed as ten-second averages plotted
every thirty-seconds for METRAC flight MFl. The deviations of the computed
balloon vertical velocity from the mean are very reasonable for real atmospheric
vertical velocities.
The results of the Minneapolis field tests demonstrate that the METRAC system
is capable of providing accurate winds on the micro and mesoscales, as well as
on the synoptic scale. This provides a single flexible system which could
be used for all atmospheric measurements such as, for example, diffusion and
turbulence studies, regional air pollution programs, and synoptic analysis.
Its application to such uses is anticipated.
-------
-113-
(LU)
O
o
O
ro
1 H 9
O
O
O
CM
I 3 H
O
O
O
I
-------
-114-
VIII. BIBLIOGRAPHY
Ackerman, B., 1974: Wind fields over the St. Louis metropolitan area,
J. Air. Poll. Cont. Assoc., 24, 232-236.
Danielsen, E. F. and R. T. Duquet, 1967: A comparison of FPS-16 and
GMD-1 measurements and methods for processing wind data, J. Appl.
Meteor.. 6, 824-836.
DeMandel, R. F. and S. J. Krivo, 1972: Measurement of small-scale
turbulent motions between the surface and 5 kilometers with the
FPS-16 Radar/Jimsphere system, preprints of Int. Conf. on
Aerospace and Aeronautical Meteor., Washington, D.C., 93-96.
Glish, John, 1971: A closed-form solution for Doppler satellite navigation,
IEEE Trans. Aerospace and Electronic Systems, AES-7, 875-878.
Kanter, I, 1962: A hyperbolic Doppler navigation system using a satellite
as a reference, Proc. East Coast Conf. on Aerospace and Navigational
Electronics, Baltimore, Md.
MITRE Corporation, 1969: Report of Trade-Off Analysis on SESAME Candidates;
Technical Report No. MTR 7013, Report under Contract No. E-27-68 (N),
Weather Bureau, ESSA.
Newton, Robert R., 1967: Everyman's Doppler satellite navigation system,
IEEE Trans. Aerospace and Electronic Systems. AES-3, 527-554.
-------
-115-
Razin, S., 1967: Explicit (non-iterative) LORAN solution, J. Inst. Navigation
Vol. 14. #3.
Schmidt, Ralph 0., 1972: A new approach to geometry of range difference location,
IEEE Trans, on Aerospace and Electronis Systems, AES-8, 821-835.
Stanford Research Institute, 1972; Regional Air Pollution Study: A Prospectus,
Final Report Contract No. 68-02-0207, Environmental Protection Agency.
------- |